Compendium Solutions NLcontrol

Exercises in
Nonlinear Control Systems

Mikael Johansson
Bo Bernhardsson
Karl Henrik Johansson
Created: 1998
Latest update: December 8, 2020
1
Introduction
The exercises are divided into problem areas that roughly match the lecture
schedule. Exercises marked “PhD” are harder than the rest. Some exercises
require a computer with software such as Matlab and Simulink.
Many people have contributed to the material in this compendium. Apart
from the authors, exercises have been suggested by Lennart Andersson, An-
ders Robertsson and Magnus Gäfvert. Exercises have also shamelessly been
borrowed (=stolen) from other sources, mainly from Karl Johan Åström’s
compendium in Nonlinear Control.
The authors, Jan 1999
Exercises marked with (H) have hints available, listed in the end of each
chapter.
2
1. Nonlinear Models and
Simulation
Exercise 1.1[Khalil, 1996]
The nonlinear dynamic equation for a pendulum is given by
mlθ¨ = −m sin θ − klθ˙,
where l > 0 is the length of the pendulum, m > 0 is the mass, k > 0 is a
friction parameter and θ is the angle subtended by the rod and the vertical
axis through the pivot point, see Figure 1.1.
Figure 1.1 The pendulum in Exercise 1.1
(a) Choose appropriate state variables and write down the state equations.
(b) Find all equilibria of the system.
(c) Linearize the system around the equilibrium points, and determine if
the system equilibria are locally asymptotically stable.

The nonlinear dynamic equations for a single-link manipulator, see Fig-
ure 1.2, with flexible joints, damping ignored, is given by
I q̈ 1 + M sin q 1 + k(q 1 − q 2 ) = 0
J q̈ 2 − k(q 1 − q 2 ) = u,
where q 1 and q 2 are angular positions, I and J are moments of inertia, k is a

spring constant, M is the total mass, L is a distance, and u is a torque input.
Choose state variables for this system and write down the state equations.
3
Chapter 1. Nonlinear Models and Simulation
q1
Figure 1.2 The flexible manipulator in Exercise 1.2

A synchronous generator connected to a strong electrical bus can be modeled
by
M δ¨ = P − Dδ˙ − η 1 E q sin δ
τ Ė q = −η 2 E q + η 3 cos δ + E F D ,
where δ is the rotor deflection angle in radians, E q is voltage, P is mechanical

input power, E F D is field voltage, D is damping coefficient, M is inertial
coefficient, τ is a time constant, and η 1 , η 2 , and η 3 are constant parameters.
(a) Using E F D as input signal and δ , δ˙, and E q as state variables, find
the state equation.
(b) Suppose that τ is relatively large so that Ė q ( 0. Show that assuming
E q to be constant reduces the model to a pendulum equation with input
torque.
(c) For the simplified model, derived in (b), find all equilibrium points.
Exercise 1.4
ψ ( t, ·)
−
r u y
C (s I − A)−1 B
Figure 1.3 The feedback system in Exercise 1.4
Figure 1.3 shows a feedback connection of a linear time-invariant system and

a nonlinear time-varying element. The variables r, u and y are vectors of the
same dimension, and ψ ( t, y) is a vector-valued function.
(a) Find a state-space model with r as input and y as output.
(b) Rewrite the pendulum model from Exercise 1.1 into the feedback con-
nection form described above.
4
+
θi sin(·) G(s) y
−
θ0 1
s
Figure 1.4 The phase-locked loop in Exercise 1.5
A phase-locked loop can be represented by the block diagram of Figure 1.4.

Let { A, B, C } be a state-space representation of the transfer function G(s).
Assume that all eigenvalues of A have negative real parts, G(0) ,= 0 and
that θ i is constant. Let z be the state of the realization { A, B, C }.
(a) Show that
ż = Az + B sin e
ė = −C z
is a state equation for the closed-loop system.

(b) Find all equilibrium points of the system.
(c) If G(s) = 1/(τ s + 1), the closed-loop model coincides with the model
of a pendulum with certain conditions on m, l, k, and (as given in
Exercise 1.1), what conditions?
Exercise 1.6 (H)
Friction
F
−
xr u x
Σ GP I D Σ 1
ms
1
s
v
−
Figure 1.5 Control system with friction in Example 1.6.
Figure 1.5 shows a block diagram of a mechanical system with friction under
PID control. The friction block is given by
F (v) = F0 sign(v)
Let xr = 0 and rewrite the system equations into feedback connection form
(i.e. a linear system in feedback with a nonlinear system).
5
Exercise 1.7
G aw
- +
r u v y
Gf f + Gp
-
Gf b
Figure 1.6 Anti-windup compensation in Example 1.7.
Figure 1.6 illustrates one approach to avoid integrator windup. Rewrite the
system into feedback connection form.
Exercise 1.8
Consider the model of a motor with a nonlinear valve in Figure 1.7. Assume
Motor Valve Process

r 1 y
Σ 1
s (s+1)2
−1
Figure 1.7 Block diagram of system in Exercise 1.8.
that the valve characteristic is given by f ( x) = x2 (this is unrealistic for

x < 0).
(a) Choose appropriate state variables and write down the state equations.
(b) For which constant input amplitudes r > 0 does the system have a
locally stable equilibrium?
(c) What would be a more realistic valve model for x < 0?
Exercise 1.9
Is the following system (a controlled nonlinear spring) nonlinear locally con-
trollable around x = ẋ = u = 0?
ẍ = − k 1 x − k 2 x3 + u.
6
Exercise 1.10Phd
The equations for the unicycle in Figure 1.8 are given by
θ
( x, y)
Figure 1.8 The “unicycle used in Exercise 1.10.
ẋ = u1 cos θ
ẏ = u1 sin θ
θ˙ = u2 ,
where ( x, y) is the position and θ the angle of the wheel. Is the system nonlin-
ear locally controllable at (0, 0, 0)? (Hint: Linearization gives no information;
use the definition directly).
Exercise 1.11PhD
The system in Figure 1.9 is known as the “rolling penny”. The equations are
θ
( x, y)
Figure 1.9 The “rolling penny” used in Exercise 1.11.
given by
ẋ = u1 cos θ
ẏ = u1 sin θ
θ˙ = u2
Ψ̇ = u1 .
Is the system nonlinear locally controllable at (0, 0, 0, 0)?
7
Exercise 1.12
Determine if the following system is nonlinear locally controllable at ( x0 , u0 ) =
(0, 0)
ẋ1 = cos( x1 ) − 1 + x22 + u

ẋ2 = sin( x1 ) + u2 .
Exercise 1.13
Simulate the system G(s) = 1/(s + 1) with a sinusoidal input u = sin ω t.
Find the amplitude of the stationary
√ output for ω = 0.5, 1, 2. Compare with
the theoretical value pG(iω )p = 1/ 1 + ω 2 .
Exercise 1.14
Consider the pendulum model given in Exercise 1.1.
(a) Make a simulation model of the system in Simulink, using for instance
m = 1, = 10, l = 1, k = 0.1. Simulate the system from various initial
states. Is the system stable? Is the equilibrium point unique? Explain
the physical intuition behind your findings.
(b) Use the function linmod in Matlab to find the linearized models for the
equilibrium points. Compare with the linearizations that you derived
in Exercise 1.1.
(c) Use a phase plane tool (such as pplane or pptool, links at the course
homepage) to construct the phase plane of the system. Compare with
the results from (a).
Exercise 1.15
Simulate the example from the lecture with two tanks, using the models
ḣ = (u − q)/ A
p √
q=a 2 h,
where h is the liquid level, u is the inflow to the tank, q the outflow, A the
cross section area of the tank, a the area of the outflow and the acceleration
due to gravity, see Figure 1.10. Use a step input flow. Make a step change
in u from u = 0 to u = c, where c is chosen in order to give a stationary
value of the heights, h 1 = h 2 = 0.1. Make a step change from u = c to
u = 0. Is the process linear? Linearize the system around h 1 = h 2 = 0.1.
Use A1 = A2 = 3 $ 10−3 , a 1 = a 2 = 7 $ 10−6 .
8
1 q
1/A f(u) 1
1 s
q q In
In Sum Gain Integrator Fcn h 1
1 In
h Out
2 In Subsystem2
h Subsystem
Figure 1.10 The flow system in Exercise 1.15
Exercise 1.16
Simulate the system with the the oscillating pivot point (the “electric hand-
saw”), see Figure 1.11. Use the equation
Figure 1.11 The inverted pendulum on an electric hand-saw in Exercise 1.16
1
θ¨ ( t) = ( + aω 2 sin ω t) sin θ ( t).
l
Assume a = 0.02m and ω = 2π · 50 for a hand-saw. Use simulation to find for
what length l the system is locally stable around θ = θ˙ = 0 (Note: asymptotic
stability is not required).
Exercise 1.17
The Lorentz equations
d
x1 = σ ( x2 − x1 )
dt
d
x2 = rx1 − x2 − x1 x3
dt
d
x3 = x1 x2 − bx3 , σ , r, b > 0,
dt
where σ , r, b are constants, are often used as example of chaotic motion.
(a) Determine all equilibrium points.
(b) Linearize the equations around x = 0 and determine for what σ , r, b
this equilibrium is locally asymptotically stable.
9
Hints
Hints
Exercise 1.6
The nonlinear system in feedback with the friction block takes − F as input
and produces V . To find the linear system, treat − F as input and V as
output.
10
2. Linearization and
Phase-Plane Analysis
Exercise 2.1[Khalil, 1996] (H)
For each of the following systems, find and classify all equilibrium points.
(a) ẋ1 = x2
ẋ2 = − x1 + x31 /6 − x2
(b) ẋ1 = − x1 + x2
ẋ2 = 0.1x1 − 2x2 − x21 − 0.1x31
(c) ẋ1 = (1 − x1 ) x1 − 2x1 x2 /(1 + x1 )

x2
ẋ2 = (1 − ) x2
1 + x1
(d) ẋ1 = x2
ẋ2 = − x1 + x2 (1 − 3x21 − 2x22 )
(e) ẋ1 = − x1 + x2 (1 + x1 )
ẋ2 = − x1 (1 + x1 )
(f) ẋ1 = ( x1 − x2 )( x21 + x22 − 1)

ẋ2 = ( x1 + x2 )( x21 + x22 − 1)
Exercise 2.2[Åström, 1968]

For all positive values of a, b and c, determine the equilibrium points of the
system
ẋ1 = ax1 − x1 x2
ẋ2 = bx21 − cx2
and determine the type of equilibrium.
11
Chapter 2. Linearization and Phase-Plane Analysis

For each of the following systems, construct the phase portrait, preferably
using a computer program, and discuss the qualitative behaviour of the
system.
(a) ẋ1 = x2
ẋ2 = x1 − 2 tan−1 ( x1 + x2 )
(b) ẋ1 = x2
ẋ2 = − x1 + x2 (1 − 3x21 − 2x22 )
(c) ẋ1 = 2x1 − x1 x2

ẋ2 = 2x21 − x2
Exercise 2.4
Saturations constitute a severe restriction for stabilization of system. Fig-
ure 2.1 shows three phase portraits, each corresponding to one of the follow-
ing linear systems under saturated feedback control.
(a) ẋ1 = x2
ẋ2 = x1 + x2 − sat(2x1 + 2x2 )
(b) ẋ1 = x2
ẋ2 = − x1 + 2x2 − sat(3x2 )
(c) ẋ1 = x2
ẋ2 = −2x1 − 2x2 − sat(− x1 − x2 )
Which phase portrait belongs to what system?
Phase plane Phase plane Phase plane

2 1 2
0.8
1.5 1.5
0.6
1 1
0.4
0.5 0.5
0.2
x2
x2
x2
0 0 0
−0.2
−0.5 −0.5
−0.4
−1 −1
−0.6
−1.5 −1.5
−0.8
−2 −1
−2 −1.5 −1 −0.5 0 0.5 1 1.5 2 −1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1 −2 −1.5 −1 −0.5 0 0.5 1 1.5 2
x1 x1 x1
Figure 2.1 Phase portraits for saturated linear systems in Exercise 2.4
12
Exercise 2.5[Khalil, 1996] (H)

The phase portraits of the following two systems are shown in Figures 2.2(a),
and 2.2(b), respectively. Mark the arrow heads and discuss the qualitative
behaviour of each system.
(a) ẋ1 = − x2
ẋ2 = x1 − x2 (1 − x21 + 0.1x41 )
(b) ẋ1 = x2
ẋ2 = x1 + x2 − 3 arctan( x1 + x2 )
Phase plane Phase plane

8 8
6 6
4 4
2 2
x2
x2
0 0
−2 −2
−4 −4
−6 −6
−8 −8
−8 −6 −4 −2 0 2 4 6 8 −8 −6 −4 −2 0 2 4 6 8
x1 x1
Figure 2.2 Phase portraits for Exercise 2.5(a) to the left, and Exercise 2.5(b) to the
right.
Exercise 2.6
The following system
ẋ1 = (u − x1 )(1 + x22 )

ẋ2 = ( x1 − 2x2 )(1 + x21 )
y = x2
is controlled by the output feedback
u = − Ky
(a) For all values of the gain K , determine the equilibrium points of the
closed loop system.
(b) Determine the equilibrium character of the origin for all values of the
parameter K . Determine in particular for what values the closed loop
system is (locally) asymptotically stable.
13

As an application of phase plane analysis, consider the model of a syn-
chronous generator derived in Exercise 1.3(b):
ẋ1 = x2
P D η1
ẋ2 = − x2 − E q sin x1 .
M M M
The equilibrium points of the system where derived in Exercise 1.3(c).Determine

the character of the equilibrium points.
Exercise 2.8 (H)

Consider the system
ẋ1 = x2 + x1 (1 − x21 − x22 )

ẋ2 = − x1 + x2 (1 − x21 − x22 )
(a) Verify that the trajectory ( x1 , x2 ) = (sin t, cos t) is a solution to the

system.
(b) The trajectory in (a) is periodic and can thus be seen as a limit cycle.
Check whether this limit cycle is stable or not.
(c) Linearize the system around the trajectory(/limit cycle).
Exercise 2.9
Linearize the ball-on-beam equation
7 2r ¨
ẍ − xφ˙2 = sin φ + φ,
5 5
around the trajectory
5 t2
3 4
sin(φ 0 ) ·
! "
φ ( t), x( t) = φ 0 ,
7 2
Exercise 2.10
Use a simple trigonometry identity to help find a nominal solution corre-
sponding to u( t) = sin (3t), y(0) = 0, ẏ(0) = 1 for the equation
4 3 1
ÿ + y ( t) = − u( t).
3 3
Linearize the equation around this nominal solution.
14
Hints
Exercise 2.11
The equations for motion of a child on a swing are given by
d d
(ml2 φ ) + m l sin φ = 0
dt dt
Here φ ( t) is the angle of the swing, m the mass, and l( t) the distance of the
child to the pivot of the swing. The child can excite the swing by changing
l( t) by moving its center of mass.
(a) Draw phase diagrams for two different constant lenghts l 1 and l 2 .
(b) Assume that it is possible to quickly change between the lenghts l 1 and
l 2 . Show how to jump between the two different systems to increase the
amplitude of the swing.
Hint: During constant l the energy in the system is constant. When l( t) changes
d
quickly φ will be continuous but dt φ ( t) will change in such a way that the
2 d
angular momentum ml dt φ is continuous.
Hints
Exercise 2.1 Set ẋ1 = ẋ2 = 0 and find necessary conditions on the stationary
points by considering the simplest equation. Use this in the other equation.
Exercise 2.5 Note that the sign of x2 determines the sign of ẋ1 .
Exercise 2.8 Introduce polar coordinates to determine stability of the limit

cycle.
x1 = r cos(θ )
x2 = r sin(θ )
with r ≥ 0.
15
3. Lyapunov Stability
Exercise 3.1
Consider the scalar system
ẋ = ax3
(a) Show that Lyapunov’s linearization method fails to determine stability

of the origin.
(b) Use the Lyapunov function
V ( x) = x4
to show that the system is globally asymptotically stable for a < 0.

(c) What can you say about the system for a = 0?
Exercise 3.2
Consider the pendulum equation with mass m and length l.
ẋ1 = x2
k
ẋ2 = − sin x1 − x2 .
l m
(a) Assume zero friction, (i.e. let k = 0), and that the mass of the pendu-
lum is concentrated at the the tip. Show that the origin is stable by
showing that the energy of the pendulum is constant along all system
trajectories.
(b) Show that the pendulum energy alone cannot be used to show asymp-
totic stability of the origin for the pendulum with non-zero friction,
k > 0. Then use LaSalle’s invariance principle to prove that the origin
is asymptotically stable.
Exercise 3.3
Consider the system
ẍ + d ẋ3 + kx = 0,
where d > 0, k > 0. Show that
1! 2
kx + ẋ2
"
V ( x, ẋ) =
2
is a Lyapunov function. Is the system locally stable, locally asymptotically
stable, and globally asymptotically stable?
16
Chapter 3. Lyapunov Stability
Exercise 3.4
Consider the linear system
−1 0
6 5
ẋ = Ax = x
1 −1
(a) Compute the eigenvalues of A and verify that the system is asymptoti-
cally stable
(b) From the lectures, we know that an equivalent characterization of
stability can be obtained by considering the Lyapunov equation
AT P + P A = − Q
where Q = QT is any positive definite matrix. The system is asymptot-

ically stable if and only if the solution P to the Lyapunov equation is
positive definite.
(i) Let
5 6
p 11 p 12
P=
p 12 p 22
Verify by completing squares that V ( x) = xT P x is a positive

definite function if and only if
p 11 > 0, p 11 p 22 − p212 > 0
(ii) Solve the Lyapunov function with Q as the identity matrix. Is the
solution P a positive definite matrix?
(c) Solve the Lyapunov equation in Matlab.
Exercise 3.5[Slotine and LI, 1991]

As you know, the system
ẋ( t) = Ax( t), t ≥ 0,
is asymptotically stable if all eigenvalues of A have negative real parts. It

might be tempted to conjecture that the time-varying system
ẋ( t) = A( t) x( t), t ≥ 0, (3.1)
is asymptotically stable if the eigenvalues of A( t) have negative real parts

for all t ≥ 0. This is not true.
(a) Show this by explicitly deriving the solution of
−1 e2t
5 6
ẋ = x, t ≥ 0.
0 −1
17
(b) The system (3.1) is however stable if the eigenvalues of A( t) + AT ( t)

have negative real parts for all t ≥ 0. Prove this by showing that
V = xT x is a Lyapunov function.
Exercise 3.6[Boyd, 1997]

A student is confronted with the nonlinear differential equation
2x
ẍ + =0
(1 + x2 )2
and is asked to determine whether or not the equation is stable. The students
think “this is an undamped mass-spring system – the spring is nonlinear with
a spring constant of 2/(1 + x2 )2 ”. The student re-writes the system as
ẋ1 = x2
−2x1
ẋ2 =
(1 + x21 )2
and constructs the obvious Lyapunov function

Z x1
2ζ 1
V ( x) = dζ + x22 .
0 (1 + ζ )
2 2 2
The student declares, “V is positive definite, because everywhere in IR2 ,

V ( x) ≥ 0, and V ( x) = 0 only if x = 0.” The student ascertains that V̇ ≤ 0
everywhere in IR2 and concludes, “the conditions for Lyapunov’s theorem are
satisfied, so the system is globally stable about x = 0”.
(a) Sadly, there is a mistake in the student’s reasoning. What is the mis-
take?
(b) Perhaps the student has merely made a poor choice of Lyapunov func-
tion, and the system really is globally stable. Is there some other Lya-
punov function that can be used to show global stability? Find such a
function, or show that no such function exists.
Exercise 3.7[Slotine and LI, 1991] (H)

Consider the system
ẋ1 = 4x21 x2 − f1 ( x1 )( x21 + 2x22 − 4)

ẋ2 = −2x31 − f2 ( x2 )( x21 + 2x22 − 4),
where the continuous functions f1 and f2 have the same sign as their argu-
ments, i.e. xi fi ( xi ) > 0 if xi ,= 0, and fi (0) = 0.
(a) Find all equilibrium points of the system. Hint: after putting the time
derivatives to zero, form a linear combination of the two equations to
conclude that either x21 + 2x22 − 4 = 0, or x1 = x2 = 0.
18
(b) Show that

E = x p x21 + 2x22 = 4
) *
is an invariant set.
(c) Show that almost all trajectories of the system tend towards the invari-
ant set E.
(d) Is E a limit cycle?
Extra: Simulate the system.
(Remark. Compare with Example 3.13 in the book by Slotine and Li.)
Exercise 3.8
Consider the system
ẋ1 = x2
ẋ2 = −2x1 − 2x2 − 4x31 .
Use the function
V ( x) = 4x21 + 2x22 + 4x41
to show that
(a) the system is globally stable around the origin.
(b) the origin is globally asymptotically stable.
Exercise 3.9
Consider the system
ÿ = sat(−3 ẏ − 2y).
(a) Show that y( t) → 0 as t → 0.
(b) For PhD students. Is it possible to prove global asymptotic stability
using a Lyapunov function V ( x) that satisfies
αpp xpp22 ≤ V ( x) ≤ β pp xpp22 , V̇ ( x) ≤ −γ pp xpp22
for some positive scalars α and β ?
(c) For PhD students. Consider the system
ẍ = u
and show that all feedback laws u = k 1 x + k 2 ẋ that give an asymptoti-
cally stable system, also give an asymptotically stable system when the
actuator saturates, i.e., when
ẍ = sat(u).
(d) For PhD students. Does the results in (c) hold for the triple integrator
d3 x
= sat(u)? (3.2)
dt3
19

The set { z ∈ IRn p V̇ ( z) = 0} arising in invariance theorems (such as LaSalle’s
theorem) is usually a ‘thin’ hypersurface, but it need not be. Carefully prove
global asymptotic stability of
ẋ1 = x2
ẋ2 = − x1 − max(0, x1 ) · max(0, x2 )
using the Lyapunov function V ( x) = xT x.
Exercise 3.11
Consider the nonlinear system
ẋ1 = − x1 + x2
ẋ2 = − x1 − x2 + ( x)
(a) Show that V ( x) = 0.5xT x is a Lyapunov function for the system when
( x) = 0.
(b) Use this Lyapunov function to show that the system is globally asymp-
totically stable for all ( x) that satisfy
( x) = ( x2 )
sign(( x2 )) = −sign( x2 )
(c) Let ( x) = x32 . This term does not satisfy the conditions in (b). However,
we can apply Lyapunov’s linearzation method to show that the origin
is still locally asymptotically stable.
For large initial values, on the other hand, simulations reveal that the
system is unstable. It would therefore be interesting to find the set of
“safe” initial values, such that all trajectories that start in this set tend
to the origin. This set is called the region of attraction of the origin. We
will now illustrate how quadratic Lyapunov functions can be used to
estimate the region of attraction.
(i) Show that V̇ ( x) < 0 for p x2 p < 1. This means that V ( x) decreases
for all solutions that are confined in the strip p x2 ( t)p < 1 for all t.
(ii) Recall that level sets for the Lyapunov function are invariant.
Thus, solutions that start inside a proper level set remain there
for all future times. Conclude that the region of attraction can be
estimated as the largest level set
Ω = { x : V ( x) ≤ γ }
for which p x2 p < 1. Compute the maxiumum value of γ .
20
Consider the second order system
ẋ1 = − x2
ẋ2 = x1 + ( x21 − 1) x2 .
(This is the Van der Pol oscillator ran backwards in time.)

(a) Use linearization and a quadratic Lyapunov function to show that the
origin is asymptotically stable.
(b) Estimate the region of attraction for the nonlinear system using the
quadratic Lyapunov derived in (a). (Hint. Transform the system into
polar coordinates, x1 = r cos(θ ), x2 = r sin(θ ) and find the largest
radius r so that Lyapunov function is decreasing for all x in the ball
Br = { x ∈ IR2 : q xq2 ≤ r}.).
(c) The procedure used in (b) tends to give quite conservative estimates of
the region of attraction. Can you think of some method to get better
estimates of the region of attraction, or to enlarge the estimate derived
above?

Consider the system
ẋ1 = x2
ẋ2 = x1 − sat(2x1 + x2 ).
(a) Show that the origin is asymptotically stable.

(b) Show that all trajectories starting in the first quadrant to the right of
the curve
x1 x2 = c
for sufficiently large c, cannot reach the origin. (Hint: Consider V ( x) =

x1 x2 ; calculate V̇ ( x) and show that on the curve V ( x) = c, the derivative
V̇ ( x) > 0 when c is sufficiently large.)
(c) Show that the origin is not globally asymptotically stable.

So far, we have only considered stability of autonomous systems, i.e. systems
without an external input. If we are faced with an open-loop unstable system
with an input that we can manipulate, a key question is whether it is possible
to find a control law that gives a stable closed-loop system. We can use
Lyapunov theory to give some answers to this question.
We say that the single-input system
ẋ = f ( x, u), x ∈ IRn , u ∈ IR
is stabilizable if there is a state feedback u = k( x) that results in a globally

asymptotically stable closed-loop system.
21
(a) Consider the special case when
ẋ = f ( x, u) = φ ( x) + ψ ( x)u,
and show that the system is stabilizable if there is positive definite

function V , such that the function
3 4 33 4 42
V V
φ ( x) − ψ ( x)
x x
is negative definite. (Hint. Try to find a feedback in terms of V ( x),

φ ( x), and ψ ( x) that makes V ( x) a Lyapunov function for the closed loop
system.)
(b) For PhD students. Show that the linear system ẋ = Ax + bu is stabi-
lizable if and only if there is a P = P T such that
A P + P AT − bbT < 0.
(Hint. Some LQR theory may come handy when proving necessity. In
particular, if the system is stabilizable, what can you say R
about the
∞
feedback law u = − kx that you obtain from the LQR cost 0 xT x +
uT u dt?)
Exercise 3.15
It can sometimes be convenient to re-write nonlinearities in a way that is
more easy to manipulate. Consider the single input, open loop stable, linear
system under saturated feedback
ẋ = Ax + Bsat(u)
u = − K x.
(a) Show that this system can be re-written in the form
ẋ = Ax + µ( x) B K x,
where 0 < µ( x) ≤ 1.
(b) Assume P > 0 is such that
xT ( AT P + P A) x ≤ 0, ∀ x
Show that all feedback gains K that satisfies
xT (( A − B K )T P + P ( A − B K )) x ≤ 0, ∀ x
guarantees the closed loop system in (a) to be stable. (The nice thing
about this formulation is that it is possible to construct efficient nu-
merical methods for simultaneously finding both feedback gains K and
Lyapunov matrix P).
22
(c) For PhD students. Consider the nonlinear system
ẋ = Ax + f ( x) + Bsat(u)
u = − K x.
Assume that the perturbation term satisfies f (0) = 0 and that f ( x) is

globally Lipschitz with Lipschitz constant k f , i.e.,
p f ( x) − f ( y)p ≤ k f p x − yp, k f > 0.
Let Q be given by the Lyapunov equation AT P + P A = − Q, with P > 0.

Show that if the Lipschitz constant satisfies
λmin ( Q)
kf < ,
2λmax ( P )
then the system is globally stabilizable by linear feedback. Also, suggest

a feedback law that stabilizes the system.
Exercise 3.16
In general, it is non-trivial to find a Lyapunov function for a given nonlinear
system. Several different methods have been derived for specific classes of
systems. In this exercise, we will investigate the following method, known as
Krasovskii’s method.
Consider systems on the form
ẋ = f ( x)
with f (0) = 0. Assume that f ( x) is continuously differentiable and that its

Jacobian, f / x, satisfies
3 4T
f f
P ( x) + ( x) P ≤ − I
x x
for all x ∈ IRn , and some matrix P = P T > 0. Then, the origin is globally
asymptotically stable with V ( x) = f T ( x) P f ( x) as Lyapunov function.
Prove the validity of the method in the following steps.
(a) Verify that f ( x) can be written as
Z1
f
f ( x) = (σ x) · x dσ .
0 x
and use this representation to show that the assumptions imply
xT P f ( x) + f T ( x) P x ≤ − xT x, ∀ x ∈ IRn
(b) Show that V ( x) = f T ( x) P f ( x) is positive definite for all x ∈ IRn .

(c) Show that V ( x) is radially unbounded.
23
Hints
(d) Using V ( x) as a Lyapunov function candidate, show that the origin is

globally asymptotically stable.
Exercise 3.17
Use Krasovskii’s method to justify Lyapunov’s linearization method.

Consider the servo system in Figure 3.18. Introduce state variables x1 and
e x2 x1
K 1
+ ( e) s+1 s
−1
Figure 3.1 The pendulum in Exercise 1.1
x2 as indicated in the figure. Assume that the reference value is zero. The
system equations can then be written as
ẋ1 = x2
ẋ2 = − x2 + K ( e) = − x2 + K (− x1 ).
Let the nonlinearity be on the form ( e) = e3 and investigate stability of the

closed loop system. (Hint: Use V ( x) = f ( x)T P f ( x) (Krassovskii’s method)
with suitable P.)
Hints
Excersice 3.7
b) Show that if x(T ) ∈ E then x21 ( t) + 2x22 ( t) = 4 for all t ≥ T.
c) Define a function V ( x) such that V = 0 on E and V ( x) > 0 if x ∈ / E, and
start by showing that V̇ ≤ 0.
24
4. Input-Output Stability
Exercise 4.1
The norms used in the definitions of stability need not be the usual Euclidian
norm. If the state-space is of finite dimension n (i.e., the state vector has n
components), stability and its type are independent of the choice of norm
(all norms are “equivalent”), although a particular choice of norm may make
analysis easier. For n = 2, draw the unit balls corresponding to the following
norms.
(a) pp xpp2 = x21 + x22 (Euclidian norm)
(b) pp xpp2 = x21 + 5x22
(c) pp xpp = p x1 p + p x2 p
(d) pp xpp = sup(p x1 p, p x2 p)
Recall that a “ball” B( x0 , R), of center x0 and radius R, is the set of x such
that pp x − x0 pp ≤ R, and that the unit ball is B(0, 1).
Exercise 4.2
r1 y1
Σ G
−
y2 r2
ψ (·) Σ
Figure 4.1 Feedback connection form
Consider an asymptotically stable linear time invariant system G intercon-

nected with a static nonlinearity ψ in the standard form (see Figure 4.1).
Compare the Nyquist, Circle, Small Gain, and Passivity Criterion with re-
spect to the following issues.
(a) What are the restrictions that must be imposed on ψ in order to apply
the different stability criteria?
(b) What restrictions must be imposed on the Nyquist curve of the linear
system in order to apply the stability criteria above?
(c) Which of the stability theorems above can handle dynamic nonlineari-
ties?
25
Chapter 4. Input-Output Stability
Exercise 4.3
1.5 1.5 1.5
1 1 1
0.5 0.5 0.5
0 0 0
−0.5 −0.5 −0.5
−1 −1 −1
−1.5 −1.5 −1.5

−2 −1.5 −1 −0.5 0 0.5 1 1.5 2 −2 −1.5 −1 −0.5 0 0.5 1 1.5 2 −2 −1.5 −1 −0.5 0 0.5 1 1.5 2
Figure 4.2 Static nonlinearities in Exercise 4.3.
Consider the static nonlinearities shown in Figure 4.2. For each nonlinearity,
(a) determine the minimal sector [α, β ],
(b) determine the gain of the nonlinearity,
(c) determine if the nonlinearity is passive.

The Nyquist curve of
4
G(s) =
(s + 1)(s/2 + 1)(s/3 + 1)
is shown in Figure 4.3 together with a circle with center in 1.5 and with
radius 2.85.
Nyquist Diagrams
1
Imaginary Axis
−1
−2
−3
−2 −1 0 1 2 3 4
Real Axis
Figure 4.3 The Nyquist-curve in Exercise 4.4
(a) Determine the maximal stability sector of the form (−α, α).
(b) Use the circle in the figure to determine another stability sector.
(c) What is the maximal stability sector of the form (0, β )?
26

The Nyquist curve of
4
G(s) =
(s − 1)(s/3 + 1)(s/5 + 1)
is shown in Figure 4.4 . Determine a possible stability sector (α, β ).
Nyquist Diagrams
1.5
0.5
Imaginary Axis
−0.5
−1
−1.5
−2
−4 −3.5 −3 −2.5 −2 −1.5 −1 −0.5 0
Real Axis
Figure 4.4 The Nyquist-curve in Exercise 4.5

Using the circle criterion, for each of the scalar transfer functions below, find
a sector (α, β ) for which the system is BIBO stable. Their Nyquist diagrams
are available in figure 4.5.
(a)
1
G(s) =
(s + 1)(s + 2)
(b)
s
G(s) =
s2 −s+1
Hint for (b): Here, G(s) is not stable. See lecture slides about the circle
criterion for an unstable system G(s).
27
0.4
0.2
Im
0
−0.2
−0.4
−0.4 −0.2 0 0.2 0.4 0.6

Re
Bode Diagram 0.8

−0.1
10
0.6
−0.3
10
Magnitude (abs)
−0.5 0.4
10
−0.7
10 0.2
−0.9
10
Im
0
270
−0.2
225
−0.4
Phase (deg)
180
−0.6
135
−0.8
90
0 1
10 10 −1
−1.5 −1 −0.5 0 0.5
Frequency (rad/sec) Re
Figure 4.5 Nyquist curves for the system in Exercise 4.6a (above) and Bode and
Nyquist curve for the system in Exercise 4.6b (below)
Exercise 4.7 (H)

Consider the linear time-varying system
ẋ( t) = ( A + Bδ ( t)C ) x,
(a) Show that the system can be written as a feedback connection of a

linear time invariant system with transfer function
G(s) = C (s I − A)−1 B
and a time-varying multiplication operator ψ defined by δ (i.e. ψ ( y) =

δ y).
(b) Let A be Hurwitz (i.e. asymptotically stable), let G(s) have one input
and one output, and let sup t≥0 pδ ( t)p ≤ 1. Show that if
sup pG(iω )p < 1

ω ∈IR
then the system is BIBO stable.

(c) Figure 4.6 shows the Nyquist curves for different transfer functions
G(s). Which transfer functions will give a BIBO stable closed loop
according to the criteria in (b)?
28
1.5 1.5 1.5
1 1 1
0.5 0.5 0.5
0 0 0
−0.5 −0.5 −0.5
−1 −1 −1
−1.5 −1.5 −1.5

−1.5 −1 −0.5 0 0.5 1 1.5 −1.5 −1 −0.5 0 0.5 1 1.5 −1.5 −1 −0.5 0 0.5 1 1.5
Figure 4.6 Nyquist curves for transfer function G(s) in Exercise 4.7.
(d) For PhD students. Let G(s) be a transfer function matrix with m in-
puts and n outputs. Show that if A is Hurwitz, pp ∆ ( t)pp ≤ 1 ∀ t, and
supω ∈IR σmax [C ( jω I − A)−1 B] < 1, then the system is BIBO stable.
Exercise 4.8
The singular values of a matrix A are denoted σi ( A).
(a) Use Matlab to compute σ ( A) for
1 10
5 6
A= .
0 1
(b) The maximal singular value is defined by
p Axp
σ1 ( A) = sup .
x p xp
Show that σ1 ( AB) ≤ σ1 ( A)σ1 ( B).
Exercise 4.9
In the previous chapter, we have seen how we can use Lyapunov functions to
prove stability of systems. In this exercise, we shall see how another type of
auxiliary functions, called storage functions, can be used to assess passivity
of a system.
Consider the nonlinear system
ẋ = f ( x, u)
y = ( x, u) (4.1)
with zero initial conditions, x(0) = 0. Show that if we can find a storage
function V ( x, u) with the following properties
• V ( x, u) is continuously differentiable.
• V (0) = 0 and V ( x, u) ≥ 0 for x ,= 0.
• uT y ≥ V̇ ( x, u).
then, the system (4.1) is passive.
29
Exercise 4.10
Let P be the solution to
AT P + P A = − I,
where A is an asymptotically stable matrix. Show that G(s) = BT P (s I −
A)−1 B is passive. (Hint. Use the function V ( x) = xT P x.)

Consider the dynamic controller
ẏ = −2y + sat( y) + u, y(0) = 0.
(a) Show that the system is passive.

(b) Is the system strictly passive?
(c) A DC motor is characterized by
θ˙ = ω
ω̇ = −ω + η,
where θ is the shaft angle and η is the input voltage. The dynamic
controller
ż = 2(θ − z) − sat(θ − z)
η = z − 2θ
is used to control the shaft position. Use any method you like to prove
that θ ( t) and ω ( t) converge to zero as t → ∞.
Exercise 4.12
(a) Let uc ( t) be an arbitrary function of time and let H (·) be a passive
system. Show that
y( t) = uc ( t) · H (uc ( t)u( t))
is passive from u to y.
(b) Show that the following adaptive system is stable
e( t) = G(s) θ ( t) − θ 0 uc ( t)
! "
θ˙ ( t) = −γ uc ( t) e( t),
if γ > 0 and G(s) is strictly passive.
Exercise 4.13PhD
Let f be a static nonlinearity in the sector (0, ∞).
30
(a) Show that the system ẋ = γ x + e, y = f ( x) is passive from e to y if

γ ≥ 0.
(b) Show that if the Popov criterion
Re (1 + γ iw)G(iω ) > 0, ∀ω,
is satisfied, with γ ≥ 0, then the closed loop system in Figure 4.7 is

absolutely stable.
1
1+γ s f (·)
1 + γs G(s)
Figure 4.7 Proof of the Popov criterion in Exercise 4.13.
(c) How does the Popov criterion change if f is in the sector (α, β ) instead?
(d) Figure 4.8 shows the Nyquist curve and the Popov curve (Re G(iω ), ω Im G(iω ))
for the system
s+1
G(s) = .
s(s + 0.1)(s2 + 0.5s + 9)
Determine a stability sector (0, β ) using the line in the figure.
0.5
−0.5
−1
−1.5
−2
−10 −9 −8 −7 −6 −5 −4 −3 −2 −1 0
Figure 4.8 Nyquist (dash-dot) and Popov curve (solid) for the system in Exer-
cise 4.13d. The Popov curve is to the right of the dashed line for all ω .
31
Hints
Hints
Exercise 4.7
b) Use the definition of L 2 -norm in the lecture slides to show that γ (ψ ) ≤ 1
by showing
qψ ( y)q2 ≤ qδ q∞ q yq2 ≤ q yq2
and then apply the appropriate theorem.
32
5. Describing Function
Analysis, Limit Cycles
Exercise 5.1 (H)
Match each of the odd, static nonlinearities in Figure 5.1 with one of the
describing functions in Figure 5.2.
Nonlinearity a b
15 15
10 10
5 5
0 0
0 5 10 0 5 10
c d
15 15
10 10
5 5
0 0
0 5 10 0 5 10
Figure 5.1 Nonlinearities in Exercise 5.1
Describing functions nr 1 nr 2
4 4
3 3
2 2
1 1
0 0
0 5 10 0 5 10
nr 3 nr 4
4 4
3 3
2 2
1 1
0 0
0 5 10 0 5 10
Figure 5.2 Describing functions N ( A) as a function of A for the odd, static nonlin-
earities in Exercise 5.1.
33
Chapter 5. Describing Function Analysis, Limit Cycles
Exercise 5.2
Compute the describing functions for
(a) the saturation,
(b) the deadzone, and
(c) the piece-wise linear function
in Figure 5.3. (Hint: Use (a) in (b) and (c).)
H H β
−D −D −D α
D D 2D α D
−H −H β
Figure 5.3 The static nonlinearities in Exercise 5.2
Exercise 5.3
Show that the describing function for a relay with hysteresis in Figure 5.4
satisfies A 
3 42 B1/2
1 πA D D
− =− 1− +i .
N ( A) 4H A A
H Im
Re
πA
4H
−D D −π D
4H
−H − N (1A)
Figure 5.4 The relay with hysteresis in Exercise 5.3.
Exercise 5.4
If the describing function for the static nonlinearity f ( x) is YN (C ), then
show that the describing function for D f ( x/ D) equals YN (C / D), where D is
a constant.
34
Exercise 5.5
Show that all odd, static nonlinearities f such that
d f ( x) d 2 f ( x)
> 0, > 0,
dx d x2
for x > 0, have a real describing function Ψ(·) that satisfies the inequalities
f ( a)
Ψ( a) < , a > 0.
a

Compute the describing function for a static nonlinearity of the form
f ( x) = k 1 x + k 2 x2 + k 3 x3 .
How does the term k 2 x2 influence the analysis?
Exercise 5.7[Slotine and LI, 1991] (H)

Consider the system in Figure 5.5, which is typical of the dynamics of elec-
tronic oscillators used in laboratories. Let
−5s
G(s) =
s2 + s + 25
0 1 y
Σ G(s)
1
Figure 5.5 Electronic oscillator.
(a) Assess intuitively the possibility of a limit cycle, by assuming that the
system is started at some small initial state, and notice that the system
can neither stay small (because of instability) nor at saturation values
(by applying the final value theorem of linear control).
(b) Use the describing function method to predict whether the system
exhibits a limit cycle. In such cases, determine the frequency and am-
plitude of the limit cycle. The describing function of a saturation is
plotted in Figure 5.6.
(c) Use the extended Nyquist criterion to assess whether the limit cycle is
stable or unstable.
35
0.8
0.6
N(A)
0.4
0.2
0
0 2 4 6 8 10
A
Figure 5.6 Normalized describing function.
Exercise 5.8
Consider a servo motor with transfer function
4
G0 (s) =
s(s + 1)(s + 2)
controlled by a relay with a dead-zone a as shown in Figure 5.7.
1
+ −a G (s)
+ −1 a 0
−
Figure 5.7 Position servo in Exercise 5.8.
(a) Show that the describing function for the relay with dead-zone a is
given by

0
 A<a
N ( A) = r
 4

1−
a2
A≥a
πA A2
Hint: √
cos(arcsin( x)) = 1 − x, for x ∈ [−1, 1]
(b) How should the parameter a be chosen so that the describing function
method predicts that sustained oscillations are avoided in the closed
loop system?
Exercise 5.9
The Ziegler-Nichols frequency response method suggest PID parameters
based on a system’s ultimate gain K u and ultimate period Tu according
to the following table. The method provides a convenient method for tuning
PID controllers, since K u and Tu can be estimated through simple experi-
ments. Once K u and Tu have been determined, the controller parameters are
directly given by the formulas above.
36
Parameter Value
K 0.6 K u
Ti 0.5Tu
Td 0.125Tu
Table 5.1 Tuning rules for Ziegler-Nichol’s method.
r e u y
G(s)
−
Figure 5.8 An auto-tuning experiment: linear system under relay feedback.
(a) Show that the parameters K u and Tu can be determined from the sus-
tained oscillations that may occur in the process under relay feedback.
Use the describing function method to give a formula for computing K u
and Tu based on oscillation data. (amplitude A and angular frequency
ω of the oscillation). Let the relay amplitude be D.
Recall that the ultimate gain and ultimate period are defined in the
following way. Let G(s) be the systems transfer function, and ω u be the
frequency where the system transfer function has a phase lag of −180
degrees. Then we have
Tu = 2π /ω u
K u = 1/pG(iω u )p
(b) What parameters would the relay method give for the process
50
G(s) =
s(s + 1)(s + 10)
which is simulated in Figure 5.9 with D = 1? Compare what you obtain
from analytical computations ( K u = 2.20, Tu = 1.99)
0.5
−0.5
−1
0 5 10 15
Figure 5.9 Input and output of system under relay feedback.
37
Exercise 5.10[Slotine and LI, 1991]

In many cases, it is desirable to limit the high frequency content in a signal.
Usually, such filtering is performed using a linear low-pass filter. Unfortu-
nately, this type of filtering introduces phase lags. If the limiter is positioned
within a feedback loop, this phase lag can be harmful for system stability.
r k y
τ1 s+1 τ2 s+1
τ2 s+1 τ1 s+1
α
τ1 > τ2
Figure 5.10 The nonlinear lowpass filter suggested in Exercise 5.10.
Figure 5.10 shows an alternative way of limiting the high frequency

content of a signal. The system is composed of a high pass filter, a saturation,
and a lowpass filter. Show that the system can be viewed as a nonlinear
lowpass filter that attenuates high-frequency inputs without introducing a
phase lag.

Consider the second order system
ẋ1 = x1 − x2 − x1 ( x21 + x22 )

ẋ2 = x1 + x2 − x2 ( x21 + x22 )
(a) Show that the unit circle is a periodic orbit.

(b) Use a Poincaré map to show that the periodic orbit derived in (a) is
asymptotically stable.
(b) Use the Lyapunov function candidate V = r2 − 1 (together with La
Salle’s Theorem) and show that the limit cycle derived in (a) is globally
Exercise 5.12PhD
Show that the system
1
G(s) =
s(s + 1)2
with relay feedback has a locally stable limit cycle.
38
Exercise 5.13PhD
Consider a linear system with relay feedback:
ẋ = Ax + Bu,
y = C x,
u = −sgn y,
where A is assumed to be non-singular. In this exercise we will study limit

cycles for this system. The Jacobian of the Poincaré map will be derived. It
gives a condition for local stability of the limit cycles.
(a) Draw a figure that illustrates a limit cycle for the system if the linear
dynamics is of third order.
(b) Consider an initial point x(0) = z that satisfies C z = 0, that is, z lies
in the switch plane. Show that the Poincaré map from z to next switch
plane intersection is
( z) = e Ah( z) z − ( e Ah( z) − I ) A−1 B, (5.1)
where h( z) is the time it takes from z to next switch plane intersection.

(c) A limit cycle corresponds to a fix point z∗ of − , that is, a point such
that z∗ = −( z∗ ). Let h( z) in (5.1) be equal to h and solve the equation
z∗ = −( z∗ ). (The solution is a function of h.)
(d) Consider a perturbed initial point z + δ z and a perturbed switch time
h + δ h. Derive a first-order Taylor expansion of and show that
( z + δ z) = − z + e Ahδ z − ( Az + B)δ h + O(δ 2 ), (5.2)
where O(δ 2 ) represents second-order and higher-order terms in δ z and

δ h.
(e) Let the perturbations δ z and δ h be such that both the perturbed initial
point and the Poincaré mapping of it lie in the switch plane. This is
equivalent to that Cδ z = 0 and C ( z + δ z) = 0. Show by using (5.2)
and C ( z + δ z) = 0 that asymptotically (when δ z → 0)
C e Ah
δh = δ z.
C ( Az + B)
(f) Use (d) and (e) to show that
( Az + B)C Ah
( z + δ z) = − z + ( I − ) e δ z.
C ( Az + B)
We have now shown that the Jacobian of the Poincaré map for a linear
system with relay feedback is equal to the matrix
( Az + B)C Ah
e .
C ( Az + B)
The limit cycle is locally stable if and only if this matrix has all eigen-
values in the unit disc.
39
Hints
(g) [Matlab exercise] Choose a state-space representation so that
1
C (s I − A)−1 B = .
(s + 1)3
In c you derived the solution z∗ = z∗ ( h). Plot the scalar function C z∗ ( h)

as a function of h. The zero crossing at h > 0 indicate a possible stable
limit cycle. What is h∗ ?
Let z = z∗ ( h∗ ) and h = h∗ . Derive the eigenvalues of
( Az + B)C Ah
e .
C ( Az + B)
Hints
Exercise 5.1
Use the interpretation of the describing function N ( A) as “equivalent gain”
for sinusoidal inputs with amplitude A.
Exercise 5.7
Q(s)
b and c) Given G(s) = P (s) , you can split the frequency response into a real
part and imaginary part as:
Q(iω ) Re{Q(iω ) P (−iω )} Im{Q(iω ) P (−iω )}

G(iω ) = = +
P (iω ) p P (iω )p2 p P (iω )p2
This is useful for plotting the Nyquist curve.
40
6. Anti-windup, Friction,
Backlash, Quantization
Exercise 6.1
Figure 6.1 (a) shows a controller in polynomial form, R(s)u = T (s)uc −
S (s) y, where u is the control signal, y the measurement variable, and uc
the reference signal. Figure (b) shows an antiwindup scheme for the same
controller. Assume that the anti-windup controller is controlling a process
given by the transfer function A(s) y = B(s)u. Also, put uc = 0.
(a) (b)
uc
uc T
T
1 u y 1 v u
∑ −S ∑
R A aw
y
−S
A aw− R
Figure 6.1 Anti-windup scheme considered in Problem 6.1
Make a block transformation to write the system in standard feedback

form with lower block P = AARA+aw
BS
− 1. Use the circle criterion to conclude that
the system is globally asymptotically stable if A is stable and the following
condition holds: 3 4
AR + BS
Re (iω ) ≥ ε > 0, ∀ω.
A A aw
Exercise 6.2
The following model for friction is described in a PhD thesis by Henrik Olsson:
dz pvp
= v− z
dt (v)
dz
F = σ0 z + σ1 (v) + Fv v,
dt
where σ0 , Fv are positive constants and (v) and σ1 (v) are positive functions
of velocity.
(a) For non-zero constant velocity, determine the stationary value of z and
its stability.
(b) What friction force does the model give in stationarity for non-zero
constant velocity?
(c) Prove that if 0 < (v) ≤ a and p z(0)p ≤ a then
p z( t)p ≤ a, t≥0
(Hint: Use the function V ( z) = z2 )
41
Chapter 6. Anti-windup, Friction, Backlash, Quantization
(d) Prove that the map from v to z is passive if z(0) = 0.

(e) Prove that the map from v to F is passive if z(0) = 0 and 0 ≤ σ1 (v) ≤
4σ0 (v)/pvp.
Exercise 6.3
Derive the describing function (v input, F output) for
(a) Coulomb friction, F = F0 sign (v)
(b) Coulomb + linear viscous friction F = F0 sign (v) + Fv v
(c) as in b) but with stiction for v = 0.
Exercise 6.4
In Lecture 7 we have studied an adaptive friction compensation scheme for
the process (assuming m = 1)
ẋ = v
v̇ = − F + u
The friction force F is given by:
F = sign(v).
If v is not directly measurable the adaptive friction compensation scheme

must be modified. Consider the following double observer scheme:
b = ( z F + K F pb
F vp)sign(b
v)
ż F b)sign(b
= − K F (u − F v)
b = zv + K v x
v
b + u − K v vb.
żv = − F
Define the estimation error states
b b
ev = v − v
b
eF = F − F b
For t such that vb( t) ,= 0 show that the state equations for the estimation
errors are given by
ḃ
ev ( t) − Kv −1 b
e v ( t)
3 4 3 43 4
=
ḃ
e F ( t) − Kv K F 0 b
e F ( t)
Conclude that the linear system is asymptotically stable if K v > 0 and

K F < 0.
42
Exercise 6.5
Figure 6.2 System considered in Problem 6.5
(a) What conclusion does describing function analysis give for the system
in Figure 6.2
(b) Show that the describing function for quantization is given by

 0
D
 A<
s 42
2
N ( A) = 4D P 2i − 1
n
3

 1− D 2n−1
< A< 2n+1
π A i=1 2A 2 D 2 D
(Hint: Use one of the nonlinearities from Lecture 6 and superposition.)
Exercise 6.6
Show that a saturation is a passive element.
Exercise 6.7
Consider the mass-spring system with dry friction
ÿ + c ẏ + ky + η( y, ẏ) = 0
where η is defined as
µ k m sign( ẏ) for p ẏp > 0

I
η( y, ẏ) =
− ky for ẏ = 0
Construct the phase potrait and discuss its qualitative behavior. (Hint: Start
by sketching the behavior for ẏ > 0 and ẏ < 0. Then discuss what happens
at ẏ = 0)
Exercise 6.8
43
+ A/D filter decim.
Figure 6.3 The static nonlinearities in Exercise 6.8
The accuracy of a crude A/D converter can be improved by adding a high-

frequency dither signal before quantization and lowpass filtering the dis-
cretized signal, see Figure 6.3. Compute the stationary value y0 of the output
if the input is a constant u0 . The dither signal is a triangle wave with zero
mean and amplitude D/2 where D is the quantization level in the A/D
converter.
Exercise 6.9
For PhD students. Show that the antiwindup scheme in observer form is
equivalent to the antiwindup scheme in polynomial form with A e equal to
the observer polynomial (see CCS for definitions).
Exercise 6.10
For PhD students. Show that the equilibrium point of an unstable linear
system preceded with a saturation can not be made globally asymptotically
stable with any control law.
44
7. Nonlinear Controller
Design
Exercise 7.1
In some cases, the main nonlinearity of a system can be isolated to a static
nonlinearity on the input. This is, for example, the case when a linear process
is controlled using a actuator with a nonlinear characteristic. A simple design
methodology is then to design a controller C (s) for the linear process and
cancel the effect of the actuator nonlinearity by feeding the computed control
through the inverse of the actuator nonlinearity, see Figure 7.1. Compute the
Controller
C (s) f −1 (·) f (·) G(s)

−
Figure 7.1 Compensation of input nonlinearity by inversion.
inverse of the following common actuator characteristics

(a) The quadratic (common in valves)
f (v) = v2 , v≥0
(b) The piecewise linear characteristic

I
k1 v pvp ≤ d
f (v) =
sign(v)( k 1 − k 2 ) d + k 2 v pvp > d
with k 1 , k 2 ≥ 0.
Use your result to derive the inverse of the important special case of a
dead zone.
(c) A backlash nonlinearity.
xout
D xin
45
Chapter 7. Nonlinear Controller Design
Exercise 7.2
An important class of nonlinear systems can be written on the form
ẋ1 = x2
ẋ2 = x3
..
.
ẋn = f ( x) + ( x)u
Assume that the full state x is available for measurement.

(a) Find a feedback
u = h( x, v)
that renders the closed loop system from the new input v to the state
linear. What conditions do you have to impose on f ( x) and ( x) in order
to make the procedure well posed?
(b) Apply this procedure to design a feedback for the inverted pendulum
ẋ1 = x2
ẋ2 = a sin( x1 ) + b cos( x2 )u
that makes the closed loop system behave as a linear system with a
double pole in s = −1. Is the control well defined for all x? Can you
explain this intuitively?
(c) One drawback with the above procedure is that it is very sensitive to
modelling errors. Show that this is the case by designing a linearizing
feedback for the system
ẋ = x2 + u
that makes the closed loop system linear with a pole in −1. Apply the
suggested control to the system
ẋ = (1 + ε ) x2 + u
and show that some solutions are unbounded irrespectively of ε ,= 0.
Exercise 7.3
Consider a linear system
ẋ1 = ax2 + bu
ẋ2 = x1
with nominal parameter values a = 1, b = 1. The system equations were

obtained by linearization of a nonlinear system, which has the consequence
that the parameters a and b vary with operating region.
46
(a) One of the design parameters in the design of a sliding mode controller
is the choice of sliding set. Which of the following sliding sets will result
in a stable sliding mode for the above system?
(i) σ ( x) = 2x1 − x2
(ii) σ ( x) = x1 + 2x2
(iii) σ ( x) = x1
(b) Let the sliding mode be σ ( x) = x1 + x2 . Construct a sliding mode
controller for the system.
(c) How large variations in the parameters a and b can the controller
designed in (b) tolerate in order to still guarantee a stable closed loop
system?
Exercise 7.4
Consider concentration control for a fluid that flows through a pipe, with no
mixing, and through a tank, with perfect mixing. A schematic diagram of
the process is shown in Figure 7.2 (left). The concentration at the inlet of
the pipe is c in ( t). Let the pipe volume be Vd and let the tank volume be Vm .
Furthermore, let the flow be q and let the concentration in the tank at the
outlet be c( t). A mass balance gives
dc( t)
Vm = q(cin ( t − L) − c( t))
dt
where L = Vd /q.
Vd k
0.63k
c in
Time
Vm a
c
L T
Figure 7.2 Schematic of the concentration control system (left). Parameters in

Ziegler-Nichols step response method (right).
(a) Show that for fixed q, the system from input c in to output c can be
reprented by a linear transfer function
K
G(s) = e−sL
sT + 1
where L and T depend on q.

(b) Use Ziegler-Nichols time response method and your model knowledge
from (a) to determine a gain scheduled PI-controller from the step
response in Figure 7.3. The step response is performed for q = 1. Recall
that the Ziegler-Nichols step response method relies on the parameters
47
Step response
0.5
Amplitude
0
−0.5
−1
0 2 4 6 8 10
Time [s]
Figure 7.3 Experimental step response for q = 1.
Controller Kp Ti Td
P 1/ a
PI 0.9/ a 3L
PID 1.2/ a 2L L/2
Table 7.1 PID parameters suggested by Ziegler-Nichols step response method.
L and a = K L/T defined in Figure 7.2 (right). (The line is tangent to

the point where the step response has maximum slope).
Given these process parameters, the method suggest PID controller
gains according to Table 7.1.
Exercise 7.5 (H)

We have seen how it in many cases can be of interest to control the system
into a set, rather than to the origin. One example of this is sliding mode
control, where the system state is forced into an invariant set, chosen in such
a way that if the state is forced onto this set, the closed loop dynamics are
exponentially stable. In this example, we will use similar ideas to design a
controller that “swings” up an inverted pendulum from its stable equilibrium
(hanging downwards) to its upright position.
Let the pendulum dynamics be given by
ẋ1 = x2
m l ml
ẋ2 = − sin( x1 ) − cos( x1 )u
Jp Jp
A general hint for this exercise: Maple and Matlab Symbolic toolbox are
handy when dealing with long equations!
(a) Denote the total energy of the pendulum by E and determine the value
E 0 corresponding to the pendulum standing in the upright position.
(b) Investigate whether the control strategy
u = k( E ( x) − E 0 )sign( x2 cos( x1 ))
48
forces the value of E towards E 0 .

(c) Draw a phase portrait of the system and discuss the qualitative be-
haviour of the closed loop system. In particular, will the suggested
control stabilize the unstable equilibrium point? Use e.g. pplane in
Matlab (link at the course homepage).
Exercise 7.6 (H)

Consider the system
ẋ1 = x1 + u
ẋ2 = x1
y = x2
Show that the control law
u = −2x1 − sign( x1 + x2 )
will make σ ( x) = x1 + x2 = 0 into a sliding mode. Determine the equivalent

dynamics on the sliding plane σ ( x) = 0.
Exercise 7.7
R1
Minimize 0
x2 ( t) + u2 ( t) dt when
ẋ( t) = u( t)
x(0) = 1
x(1) = 0
Exercise 7.8
Neglecting air resistance and the curvature of the earth the launching of a
satellite is described with the following equations
ẋ1 = x3
ẋ2 = x4
F
ẋ3 = cos u
m
F
ẋ4 = sin u −
m
Here x1 is the horizontal and x2 the vertical coordinate and x3 and x4 are the
corresponding velocities. The signal u is the controlled angle. The criterion
is to maximize 0.1x1 + x2 + 5x3 + 3x4 at the end point. Show that the optimal
control signal has the form
At + B
tan u =
Ct + D
and determine A, B, C, D.
49
Exercise 7.9
Suppose more realistically that m and F vary. Let F = u2 ( t) be a control
signal with limitations
0 ≤ u2 ( t) ≤ umax
and let the mass m = x5 ( t) vary as
ẋ5 = −γ u2
Show that
 
 λ4 
 umax σ <0
u2 > 0
tan u1 = λ3 and u2 = 0 σ >0
 

⋆ u2 = 0 ⋆ σ = 0,
where ⋆ means that the solution is unknown. Determine equations for λ and
σ . (You do not have to solve these equations).
Exercise 7.10
Consider the system
I
ẋ1 = x2
ẋ2 = − x1 − x32 + (1 + x1 )u
with initial conditions x1 (0) = 1, x2 (0) = 1 and let the criterion be

Z1
e x1 + x22 + u2 dt.
2
min
0
Is the problem normal ? Show that extremals satisfy
ẋ1 = f1 ( x1 , x2 , λ1 , λ2 )
ẋ2 = f2 ( x1 , x2 , λ1 , λ2 )
λ̇1 = f3 ( x1 , x2 , λ1 , λ2 )
λ̇2 = f4 ( x1 , x2 , λ1 , λ2 )
Determine f1 , f2 , f3 , f4 . What conditions must λ1 , λ2 satisfy at the end point?
Exercise 7.11
Consider the double integrator
ẋ1 = x2
ẋ2 = u, pup ≤ 1
50
with initial value x(0) = x0 and input constraint −umax ≤ u ≤ umax . We

are interested in finding the control that brings the system to rest (x1 ( t f ) =
0, x2 ( t f ) = 0) in minimum time. (You may think of this as a way of designing
a controller that reacts quickly on set-point changes) Show that the optimal
control is of “bang-bang type” with at most one switch. In other words, the
control can be expressed as the feedback law
σ ( x) > 0
I
umax
u=
−umax σ ( x) < 0
Draw a phase portrait of the closed loop system under the optimal control.
Exercise 7.12
Consider the problem of controlling the double integrator
I
ẋ1 = x2
ẋ2 = u, pup ≤ 1
from an arbitrary intitial condition x(0) to the origin so that the criterion
Z tf
(1 + pup) dt
0
is minimized (t f is the first time so that x( t f ) = 0). Show that all extremals
are of the form 
 −1 0 ≤ t ≤ t 1

u( t) = 0 t1 ≤ t ≤ t2


1 t2 ≤ t ≤ t f
or 
 1
 0 ≤ t ≤ t1
u( t) = 0 t1 ≤ t ≤ t2


−1 t 2 ≤ t ≤ t f
for some t 1 , t 2 with 0 ≤ t 1 ≤ t 2 ≤ t f . Some time interval can have the length
0. Assume that the problem is normal.
Exercise 7.13
Consider the system
   
 − 5 2 0

ẋ =  
x+   u
−6 2 1
 
1
 
from x0 = 0 to x( t f ) =   in minimum time with pu( t)p ≤ 3. Show that
1
the optimal controller is either
I
−3 0 ≤ t ≤ t 1
u( t) =
+3 t 1 ≤ t ≤ t f
51
or I
+3 0 ≤ t ≤ t 1
u( t) =
−3 t 1 ≤ t ≤ t f
for some t 1 .
Exercise 7.14 (H)

Show that minimum time control of a linear system
ẋ = Ax + Bu, pup ≤ 1, x( t f ) = 0,
leads to extremals of the form
u( t) = −sign (C T e− At B)
for some vector C. What does this say about the optimal input when A =
B = 1?
Exercise 7.15
What is the conclusion from the maximum principle for the problem
Z1
min u dt,
0
ẋ1 = u
x1 (0) = 0
x1 (1) = 1
Explain.
Exercise 7.16
Consider the control system
ẍ − 2( ẋ)2 + x = u − 1 (7.1)
(a) Write the system in first-order state-space form.

(b) Suppose u( t) " 0. Find all equilibria and determine if they are sta-
ble or asymptotically stable, if possible, using Lyapunovs linearization
method.
(c) Design a state-feedback controller u = u( x, ẋ) for (7.1), such that the
origin of the closed loop system is globally asymptotically stable.
Exercise 7.17 (H)
52
This problem will use the Lyapunov method for design of a control signal
which will stabilize a system. Consider the system
ẋ1 = − x1 + x2 + x3 · tan( x1 )
ẋ2 = − x32 − x1 (7.2)
ẋ3 = x22 +u
Choose u = u( x1 , x2 , x3 ) such that the closed loop system becomes globally

Exercise 7.18 (H)

A nonlinear system is given below.
x˙1 = −3x1 + x31 − x2 + u

x˙2 = x1 − ax2
(a) Using Lyapunovs linearizatio method determine, if possible, the (local)

stability properties of all equilibrium points to the nonlinear system if
u( t) " 0 and a = 1.
(b) If u( t) " 0 and a = 0, prove that the origin is locally asymptotically
stable using the Lyapunov function candidate
1 2 1 2
V ( x) = x + x
2 1 2 2
combined with LaSalles theorem.

(c) If a = 1, determine a nonlinear state feedback control u = f ( x) such
that the origin is globally asymptotically stable.
Exercise 7.19
In this problem we are going to examine how to stabilize a system using a
bounded control signal u = sat5 (v), i.e.,

 5,
 v ≥ 5;
u(v) = v, −5 ≤ v ≤ 5;


−5, v ≤ −5;
Your task is to choose the control signal v = v( x1 , x2 ), such that the sys-
tem (7.3)
ẋ1 = x1 x2
ẋ2 = u (7.3)
u = sat5 (v)
is globally asymptotically stabilized.
53
(a) Indicate the problem of using the Lyapunov candidate
Va = x21 + x22
to design a globally stabilized system when using bounded control.

(b) Instead try with the Lyapunov function candidate
Vb = log(1 + x21 ) + x22
and choose v( x1 , x2 ) so that the system is globally asymptotically stabi-

lized.
Exercise 7.20
Consider the system
ẋ1 = x2 − x1
ẋ2 = kx21 − x2 + u (7.4)
where u is the input and k is an unknown coefficient. Since k is unknown,

we can not use the traditional methods to design a stabilizing controller. We
therefore try an approach where we estimate k (denoted k̂), and try to prove
stability for the whole system (this is often denoted adaptive control). We
assume that k is unknown, and changes very slowly.
Find an update law for k̂ and a stabilizing controller" u. Use the Lyapunov
1
! 2 2
function candidate V ( x1 , x2 , k̂) = 2 x1 + x2 + ( k − k̂)2 .
Exercise 7.21
Consider the system
ẋ1 = x21 + x2
ẋ2 = u
Compute a controller using back-stepping to globally stabilize the origin.
Exercise 7.22
Consider the system
ẋ1 = x21 − x31 + x2

ẋ2 = u
(a) Compute a controller using back-stepping to globally stabilize the ori-

gin.
b) Draw a phase plane plot using Matlab (pplane).
54
Exercise 7.23
Consider the following nonlinear system:
ẋ1 = x1 + x2
ẋ2 = sin( x1 − x2 ) + u
(a) Show that the system is on strict feedback form.

(b) Design a controller based on back-stepping for the system.
Exercise 7.24
ẋ1 = −sat( x1 ) + x21 x2

ẋ2 = x21 + u
(a) Show that the system is on strict feedback form.

(b) Design a controller based on back-stepping for the system.
Exercise 7.25
ẋ1 = x1 + x2
ẋ2 = sin( x1 − x2 ) + x3
ẋ3 = u
Design a controller based on back-stepping for the system. You do not need
to substitute back to x1 , x2 , x3 in the computed controller.
Exercise 7.26
Consider the discrete time system
x k+1 = f k ( x k , u k ), k = 0, 1, . . . , N − 1
where f k ( x, u) = 2x+u and N = 4. We want to find the optimal state-feedback

law u k = µ k ( x k ) that minimizes the cost
N
X −1
N (xN ) + k ( x k , µ k ( x k ))
k=0
where
4 ( x) = x2
3 ( x, u) = x2 + u2
2 ( x, u) = x2 + 3u2
1 ( x, u) = x2 + 7u2
0 ( x, u) = x2 + 15u2
55
Hints
(a) Use dynamic programming to compute the optimal state-feedback law

with respect to the given cost function.
(b) Compute the optimal cost and the optimal control sequence u0 , u1 , u2 , u3
when the initial state is x0 = −2.
(c) Assume in addition that the control action at time k = 0 is limited to
pu0 p ≤ 1. Compute the cost-to-go function V0 ( x0 ).
Exercise 7.27
Consider the nonlinear optimal control problem
Z1
minimize ( x( t)u( t))2 dt + x(1)2 ,
0
subject to ẋ( t) = x( t)u( t), x(0) = 1.
Solve the problem using dynamic programming by making the ansatz V ( t, x) =

q( t) x2 .
Hints
Exercise 7.5
Use a Lyapunov function argument with V ( x) = ( E ( x) − E 0 )2 .
Exercise 7.6
Use V ( x) = σ 2 ( x)/2.
Exercise 7.14
You will most likely need the following relations. If
y = eAx [ x = e− A y
and
T
( e A )T = e A
Exercise 7.17
1 2
Use V ( x1 , x2 , x3 ) = ( x + x22 + x23 ).
2 1
Exercise 7.18
Use the Lyapunov function candidate from (b).
56
Solutions to Chapter 1
Solution 1.1
(a) Choose the angular position and velocity as state variables, i.e., let
x1 = θ
x2 = θ˙
We obtain
ẋ1 = x2
k
ẋ2 = − sin( x1 ) − x2
l m
(b) By setting the state derivatives to zero, we obtain
0 = x2
k
0=− sin( x1 ) − x2
l m
We find the equilibrium points ( x1 , x2 ) = (nπ, 0) with n = 0, ±1, ±2, . . . .

The equilibrium points correspond to the pendulum hanging down (n
even), or the pendulum balancing in the upright position (n odd).
(c) Linearization gives
0 1
5 6
d
∆x = ∆x (7.5)
dt − l (−1)n − mk
The linearized system is stable for even n, and unstable for odd n. We
can use Lyapunov’s linearization method to conclude that the pendulum
is LAS around the lower equilibrium point, and unstable around the
upper equilibrium point.
Solution 1.2
We choose angular positions and velocities as state variables. Letting x1 = q 1 ,
x2 = q̇ 1 , x3 = q 2 , x4 = q̇ 2 , we obtain
ẋ1 = x2
ML k
ẋ2 = − sin x1 − ( x1 − x3 )
I I
ẋ3 = x4
k 1
ẋ4 = ( x1 − x3 ) + u
J J
57
Solution 1.3
(a) Let x1 = δ , x2 = δ˙, x3 = E q and u = E F D . We obtain
ẋ1 = x2
P D η1
ẋ2 = − x2 − x3 sin x1
M M M
η2 η3 1
ẋ3 = − x3 + cos x1 + u
τ τ τ
(b) With E q being constant, the model reduces to
ẋ1 = x2
P D η1
ẋ2 = − x2 − E q sin x1
M M M
which is the pendulum equation with input torque P / M.

(c) The equilibrium points of the simplified equation are given by sin x1 =
η 1 E q , x2 = 0.
P
Solution 1.4
(a) Let
ẋ = Ax + Bu, y = Cx
be a state-space realization of the linear system. We have
u = r − ψ ( t, y) = r − ψ ( t, C x)
and hence
ẋ = Ax − Bψ ( t, C x) + Br, y = Cx
(b) To separate the linear dynamics from the nonlinearities, write the
pendulum state equations as
ẋ1 = x2
k
ẋ2 = − x2 − sin( x1 )
m l
and view sin( x1 ) as an input, generated by a nonlinear feedback from

y = x1 (Compare with Figure 1.3). Introduce the state vector x =
( x1 , x2 )T , and re-write the equations as
0 1 0
5 6 5 6
ẋ = x+ u (7.6)
0 − k/m /l
y = [1 0] x (7.7)
u = − sin( y), (7.8)
which is on the requested form.
58
Solution 1.5
(a) Hint: ė = − y = −C z.
(b) Assume ż = ė = 0, i.e.
0 = Az + B sin( e)
0 = −C z
Since A is invertible if and only if all eigenvalues are non-zero – which

is assumed – we may solve for z in the upper equation according to
z = − A−1 B sin( e)
Replacing z with this expression in the lower equation and recalling

that G(s) = C (s I − A)−1 B, we then get
0 = −C z = −C (− A−1 B sin( e)) = −C (0 I − A)−1 B sin( e) = −G(0) sin( e)
But since G(0) was assumed to be non-zero, this means that sin( e) = 0
and hence e = nπ for any integer n. With this choice of e, the upper
equation reduces to 0 = Az and hence z = 0, since A was invertible.
What we have shown is that if there exist equilibrium points, they have
to be on the form ( z, e) = (0, nπ ). Simply verifying that ż = ė = 0
for these points shows that the equilibrium points are indeed given by
( z, e) = (0, nπ ) for any integer n!
Alternative solution: The equilibrium points are given by ż = 0 and
ė = 0. In steady-state (at an equilibrium point) the amplification of the
transfer function is G(0). Denote the steady state error eo = θ i − θ 0 . If
this should be constant it means that θ 0 is constant (see block diagram
of Fig.1.4) and thus θ˙0 = 0, which is the same signal as the output
y =[
0 = G(0) sin eo
from which we obtain
eo = ±nπ, n = 0, 1, 2, . . .
(c) For G(s) = 1/(τ s + 1), we take A = −1/τ , B = 1/τ and C = 1. Then
1 1
ż = − z + sin e
τ τ
ė = − z
Now, let x1 = e and x2 = − z, so that
ẋ1 = x2
1 1
ẋ2 = − x2 − sin x1 ,
τ τ
which is the pendulum model with /l = k/m = 1/τ .
59
Solution 1.6
Let G P I D (s) be the transfer function for the PID controller. In order to find
the feedback interconnection form, the first step is to define the input and
output of the non-linearity. In this case F is the output and v is the input.
This implies, that the required linear system Gl has as its input F and as
its output v. By the help of the block-diagram one finds
s
Gl (s) =
ms2 + G P I D (s)
Hence, the whole system is a feedback interconnection of the linear system
Gl (s) and the non-linearity F (v). Observe, the feedback interconnection form
is usually defined such that, the linear part receives a negative feedback from
the non-linearity.
Solution 1.7
The requested form
U (s) = −Gl (s) V (s)

v( t) = sat(u)
is obtained with
G f bGp − G aw
Gl (s) = .
1 + G aw
Solution 1.8
(a) Introduce x1 = y, x2 = ẏ, x3 = “valve input”, then
ẋ1 = x2
ẋ2 = −2x2 − x1 + f ( x3 )
ẋ3 = r − x1
√
(b) For a constant input r the √ equilibrium point is given by x = (r, 0, ± r).
The linearization for x = (r, 0, r) has
0 1 0
 
√
A =  −1 −2 2 r  .
−1 0 0
The characteristic equation is given by
√
λ2 (λ + 2) + 2 r + λ = 0.
The condition for stability of λ3 + √ aλ2 + bλ + c is a, b, c > 0 and ab > c.

Similiar calculations for x = (r, 0, − r) gives that that equilibrium point is
never locally stable. Hence we get local stability if 0 < r < 1. An alterna-
tive approach is to compute that the gain margin of s(s+1 1)2 is 2. Since the
√
linearization of the valve has gain f ′ ( x3 ) = 2x3 = 2 r we get stability for
r < 1.
60
(c) More realistic is that the flow is zero.
Solution 1.9
The linearization is given by
ẍ = − k 1 x + u,
which is controllable. Hence the system is nonlinear locally controllable.
Solution 1.10
The linearized system is not controllable. The system is however nonlinear
locally controllable. This can be seen directly from the definition as follows:
We must show that we can drive the system from (0, 0, 0) to a near by
state ( xT , yT , θ T ) using small control signals u1 and u2 . By the sequence
u = (u1 , u2 ) = (0, ε1 ),u = (ε1 , 0),u = (0, −ε1 ), u = (−ε2 , 0) (or in words: "turn
left, forward, turn right, backwards") one can move to the state (0, yT , 0).
Then apply (ε3 , 0) and then (0, ε4 ) to end up in ( xT , yT , θ T ). For any time
T > 0 this movement can be done with small εi if xT , yT and θ T are small.
Solution 1.11
Same solution as in 1.10, except that you have to find a movement afterwards
that changes Ψ without changing the other states. This can be done by
the sequence: L-F-R-B-R-F-L-B where F=forward, B=backwards, L=turn left,
R=turn right.
Solution 1.12
The linearized system at ( x0 , u0 ) is
ẋ1 = u
ẋ2 = x1
The controllability matrix
1 0
5 6
Wc =
0 1
has full rank. Since the linearized system is controllable the nonlinear system
is also locally controllable at ( x0 , u0 ).
Solution 1.13
See lecture slides. Why does max(abs(y(:))) not give the correct stationary
output amplitude?.
Solution 1.14
61
Figure 7.4 Solution for 1.14
(a) See Fig 7.4 for an example of an implementation in matlab. Note that
the outputs have been names x1 and x2 . This can make the linearization
in the next task be easier to interpret
(b) linmod(’pendulum’,[0,0]), if the file is named pendulum.slx. The lin-
earization should be the same as in 1.1. However x1 could either be θ
or θ˙ and vice versa for x2 .
Solution 1.15
See lecture slides.
Solution 1.16
With a = 0.02 and w = 100π we get local stability for l ∈ [0.044, 1.9].
Solution 1.17
p p
(a) x =
p 0, and if rp> 1 also x = ( b(r − 1), b(r − 1), r − 1) and x =
(− b(r − 1), − b(r − 1), r − 1).
(b) The linearization around x = 0 is
 
−σ σ 0
 
ẋ =  r −1 0 
0 0 −b
with characteristic polynomial (s + b)(s2 + (σ + 1)s + σ (1 − r). A second

order monic polynomial has all roots in the left half plane iff it has
positive coefficients, hence x = 0 is LAS when 0 < r < 1, (σ , b, r > 0 by
default).
62
Solution 2.1
(a) The equilibrium points are
√ √
( x1 , x2 ) = (0, 0), ( 6, 0), (− 6, 0),
which are stable focus, saddle, and saddle, respectively.

(b) The equilibrium points are
( x1 , x2 ) = (0, 0), (−2.5505, −2.5505), (−7.4495, −7.4495),
which are stable node, saddle point, and stable focus, respectively.
(c) The equilibrium points are
( x1 , x2 ) = (0, 0), (1, 0), (0, 1),
which are unstable node, saddle, and stable node, respectively.

(d) The equilibrium is
( x1 , x2 ) = (0, 0),
which is an unstable focus.

(e) The equilibrium point is
( x1 , x2 ) = (0, 0),
which is a stable focus.

(f) The system has an equilibrium set
x21 + x22 = 1
and an equilibrium point
( x1 , x2 ) = (0, 0),
which is a stable focus.
Solution 2.2
The three equilibrium points are
q q
( x1 , x2 ) = (0, 0), ( ( ac/b), a), (− ( ac/b), a).
The first equilibrium point is a saddle. The other equilibria are stable nodes
if 8a ≤ c and stable focuses if 8a > c.
63
Solution 2.3
(a) The system has three equilibrium points
( x1 , x2 ) = (0, 0), ( a, 0), (− a, 0)
where a is the first positive root of
a
a − tan( ) = 0
2
given, approximately, by a = 2.33. The origin is a stable node, while

the other two equilibria are saddles.
(b) The system has the origin as a unique equilibrium point, being an
unstable focus.
(c) The system has the equilibrium points
( x1 , x2 ) = (0, 0), (1, 2), (−1, 2),
which are saddle, stable focus, and stable focus, respectively.
Solution 2.4
Close to the origin, the saturation element opens in the linear region, and
all system are assigned the same closed loop dynamics. Far away from the
origin, the influence of the saturated control can be neglected, and the open
loop dynamics governs the behaviour.
(a) System (a) has one stable and one unstable eigenvalue. For initial val-
ues close to the stable eigenvector, the state will move towards the
origin. For initial values close to the unstable eigenvector, the system
diverges towards infinity. This corresponds to the rightmost phase por-
trait.
(b) All eigenvalues of system (b) are unstable. Thus, for initial values
sufficiently far from the origin, the system state will diverge. This
corresponds to the leftmost phase portrait. Note how the region of
attraction (the set of initial states, for which the state converges to the
origin) is severely limited.
(c) System (c) is stable also in open loop. This corresponds to the phase
portrait in the middle.
Solution 2.5
(a) From the state equations we see that the system has the origin as
a unique equilibrium point. To determine the direction of the arrow
heads we note that if x2 > 0 then ẋ1 < 0, and if x2 < 0 then ẋ1 > 0.
Hence, x1 moves to the left in the upper half plane, and to the right in
the lower half plane. After marking the arrow heads in the plots we see
that the origin is a stable focus. This can be determined by inspection
of the vector fields. We also see that the system has two limit cycles.
The inner one is unstable and the outer one is stable.
64
(b) The system has three equilibrium points
( x1 , x2 ) = (0, 0), ( a, 0), (− a, 0),
where we can see that a ( 4. As before we the sign of x2 determines the

sign of ẋ1 which makes marking the arrow heads an easy task. After
marking the arrow heads we note that the origin is a stable focus, and
the points ( a, 0), (− a, 0) are saddles (trajectories approach the point
from one direction and leave along another direction).
Solution 2.6
(a) The equilibrium points are obtained by setting ẋ = 0. For K ,= −2, the
origin is the unique equilibrium point. When K = −2, the line x1 = 2x2
is an equilibrium set.
(b) The Jacobian is given by
−1 − K
5 6
f
(0) =
x 1 −2
with eigenvalues
r
3 1
λ=− ± − K.
2 4
Thus, the closed loop system is asymptotically stable about the origin
for K > −2. Depending on the value of K , we can origin has the
following character
1
<K stable focus
4
1
−2 < K < stable node
4
K < −2 saddle.
Solution 2.7
The equilibria are given by sin x01 = P
η Eq , x02 = 0. The characteristic equation
for the linearization becomes
λ2 + αλ + β = 0,
> 0 and β = Mq cos x01 . Depending on α, β the equilibria are

ηE
where α = M D
stable focus, stable nodes or saddle points.
65
Solution 2.8
(a) Just plug into the system dynamics.
(b) To determine stability of the limit cycle, we introduce polar coordinates.
With r ≥ 0:
x1 = r cos(θ )
x2 = r sin(θ )
Differentiating both sides gives
cos(θ ) −r sin(θ )
3 4 3 43 4
ẋ1 ṙ
=
ẋ2 sin(θ ) r cos(θ ) θ˙
Inverting the matrix gives:
1 r cos(θ ) r sin(θ )
3 4 3 43 4
ṙ ẋ1
=
θ˙ r − sin(θ ) cos(θ ) ẋ2
Plugging in the state equations results in:
ṙ = r(1 − r2 ) (7.9)
θ˙ = −1 (7.10)
We see that the the only equilibrium points to (7.9) are 0 and 1 (since
r ≥ 0). Linearizing around r = 1 (i.e. the limit cycle) gives:
r̃˙ = −2r̃
which implies that the the r = 1 is a locally asymptotically stable equi-

librium point of (7.9). Hence the limit cycle is stable.
Alternative for determining stability of limit cycle We can also

use LaSalle’s invariance principle: Define
V ( x) = ( x21 + x22 − 1)2
Then
V̇ = . . . = −4( x21 + x22 − 1)2 ( x21 + x22 ) ≤ 0 ∀ x ∈ R2 (7.11)
Therefore
1
Ω= x : ≤ pp xpp2 ≤ 2
) *
2
is a compact invariant set. Let
E = x ∈ Ω : V̇ ( x) = 0
) *
From (7.11) we see that
E = x : pp xpp2 = 1
) *
which is the limit cycle itself. By the invariance principle we conclude

that all trajectories staring in Ω (i.e. in a neighborhood of the limit
cycle) will converge to E (the limit cycle). The limit cycle is thus stable.
66
(c) In compact notation we have:
ẋ( t) = f ( x( t))
Introduce x̃( t) = x( t) − x0 ( t) as the deviation from the nominal trajec-

tory. It holds that
ẋ( t) = ẋ0 ( t) + x̃˙ ( t).
The first order Taylor expansion of f around x0 ( t) is given by
f ( x0 ( t))
f ( x( t)) ( f ( x0 ( t)) + ( x( t) − x0 ( t))
x | {z }
x̃( t)
which gives that
f ( x0 ( t))
ẋ( t) ( f ( x0 ( t)) + x̃( t)
x
So
f ( x0 ( t))
ẋ0 ( t) + x̃˙ ( t) = f ( x0 ( t)) + x̃( t)
x
In subproblem a) we showed that x0 ( t) is a solution to the system, i.e.
ẋ0 ( t) = f ( x0 ( t)) and thus
f ( x0 ( t))
x̃˙( t) = x̃ = A( t) x̃( t)
x
where
A f1( x0 ( t)) f1 ( x0 ( t))
−2 sin2 ( t)
B
1 − sin(2t)
3 4
x1 x2
A( t) = = .
f2 ( x0 ( t)) f2 ( x0 ( t)) −1 − sin(2t) −2 cos2 ( t)
x1 x2
Solution 2.9
7 2r
x̃¨ = cos(φ 0 )φ˜ + φ¨˜
5 5
Solution 2.10
Using the identity
3 1
(sin t)3 = sin t − sin 3t
4 4
we see that u0 ( t) = sin (3t), y0 ( t) = sin t is a nominal solution. The
linearization is given by
1
ỹ¨ + 4 sin2 t · ỹ = − ũ.
3
Solution 2.11
No solution yet.
67
Solution 3.1
(a) Linearization about the system around the origin yields
f
A= = 3ax2
x
Thus, at the origin we have A = 0. Since the linearization has one
eigenvalue on the imaginary axis, linearization fails to determine sta-
bility of the origin.
(b) V (0) = 0, V ( x) ,= 0 for x ,= 0, and V ( x) → ∞ as x → ∞. Thus, V ( x)
satisfies the conditions for being a Lyapunov function candidate. Its
time derivative is
V
V̇ ( x) = f ( x) = 4ax6 (7.12)
x
which is negative definite for a < 0. The desired result now follows
from Lyapunov’s global asymptotic stability theorem.
(c) For a = 0, the system is linear and given by
ẋ = 0
The system has solutions x( t) = x0 for all t. Thus, the system is stable.
A similar conclusion can be drawn from the Lyapunov function used in
(b).
Solution 3.2
(a) Since x2 is angular velocity, the speed of the pendulum tip is given by
lx2 . Since we assume that all mass is concentrated at the tip the kinetic
energy of the pendulum is
ml2 x22
.
2
The potential energy is given by m h, where h is the vertical position of
the pendulum relative to some reference level. We choose this reference
level by letting h = 0 when x1 = 0 (i.e pendulum in downward position).
h can expressed as
h = l(1 − cos( x1 ))
The pendulum’s total energy is then given by
ml2 2
V ( x) = m l(1 − cos( x1 )) + x
2 2
We use V as a candidate Lyapunov function. We see that V is positive
for x ,= 0 and V (0) = 0, and compute the time derivative
dV ( x) X V
= ẋi = m l sin( x1 ) x2 + x2 (−m l sin( x1 )) = 0
dt xi
i
68
V is thus a Lyapunov function. From Lyapunov’s theorem, we conclude

that the origin is a stable equilibrium point. Since V̇ ( x) = 0, we can
also conclude that the origin is not asymptotically stable; trajectories
starting at a level surface V ( x) = c, remain on this surface for all
future times.
(b) For k ,= 0, using V ( x) as above, similar calculations give
dV ( x)
= − kl2 x22
dt
V̇ ( x) is negative semidefinite. It is not negative definite because V̇ ( x) =

0 for x2 = 0. In other words V̇ ( x) = 0 along the x1 axis. To show local
asymptotic stability of the origin define
E = ( x1 , x2 ) p V̇ ( x) = 0 = {( x1 , x2 ) p x2 = 0}
) *
Suppose x2 = 0, and x1 ,= 0. Then by the second state equation we have

ẋ2 = − sin x1 ,= 0, p x1 p ≤ 0.9π
l
Thus the largest invariant set in E is {0}. (Note that since we are
considering local asymptotic stability, it is sufficient to consider p x1 p ≤
π − ε for any sufficiently small positive ε.) By LaSalle’s invariance
principle we conclude that x → 0.
Solution 3.3
With V = kx2 /2 + ẋ2 /2 we get V̇ = − d ẋ4 ≤ 0. Since V̇ = 0 only when
ẋ = 0 and the system equation then gives ẍ = − kx ,= 0 unless also x = 0,
we conclude that x = ẋ = 0 is the only invariant set. The origin is globally
asymptotically stable since the Lyapunov function is radially unbounded.
Solution 3.4
√
(a) The eigenvalues of A are λ = −1/2 ± i 3/2.
(b) (i) We have
V ( x) = p 11 x21 + 2p 12 x1 x2 + p 22 x22 = (if p 11 ,= 0)

p 12 p2
= p 11 ( x1 + x2 )2 + (p 22 − 12 ) x22
p 11 p 11
If p 11 > 0 and p 11 p 22 − p212 > 0, both terms are non-negative.

Moreover, V ( x) → ∞ as x → ∞, and V ( x) = 0 [ x1 = x2 = 0
(This proves the "if"-part). If the conditions on p i j do not hold, it
is easy to find x such that V ( x) < 0 (proving the "only if"-part).
(ii) We want to solve
0 1 p 11 p 12 p 11 p 12 0 −1 −1 0
5 65 6 5 65 6 5 6
+ =
−1 −1 p 12 p 22 p 12 p 22 1 −1 0 −1
69
Reading off the elements, we see that


 2p = −1
 12

p 22 − p 11 − p 12 =0



−2p 12 − 2p 22 = −1
which has the solution p 11 = 1.5, p 12 = −0.5 and p 22 = 1. P is a

positive definite matrix.
(c) Use the Matlab command lyap(A’,eye(2)).
Solution 3.6
(a) The mistake is that V is not radially unbounded. The student has
forgotten to check that lim x→∞ V ( x) = ∞. In fact,
x21 1 2
V ( x1 , x2 ) = + x
1+ x21 2 2
so that lim x→∞ = 1. Consequently, V is not radially unbounded.

(b) No, the problem is not that the student was not clever enough to find
a Lyapunov function. There is no Lyapunov function, since the system
is not globally stable. Let’s show this now. In part (a), you may have
noticed that V̇ = 0 for all x. In other words, V is an “integral” of the
motion; the trajectories lie on the curves of constant value of V , i.e., we
have
1 2 x21
V ( x) = x2 + = V ( x0 ) = c
2 1 + x21
If c > 1 then x( t) cannot change sign, since
x21
x22 = c − ≥ c−1
1 + x21
√
In this case, we have p x2 p ≥ c − 1. Since ẋ1 = x2 , it follows that
p x1 p → ∞ as t → ∞. Roughly speaking, if the system starts with more
initial stored energy than can possibly be stored as potential energy in
the spring, the trajectories will diverge.
Solution 3.7
Find the equilibrium points for the system.
ẋ1 = 0 = 4x21 x2 − f1 ( x1 )( x21 + 2x22 − 4) (1)

ẋ2 = 0 = −2x31 − f2 ( x2 )( x21 + 2x22 − 4), (2)
70
(a) Adding equation (1) times x1 to equation (2) times 2x2 gives
− ( x1 f1 ( x1 ) + 2x2 f2 ( x2 )) ( x21 + 2x22 − 4) = 0
From this we see that either x1 = x2 = 0 or x21 + 2x22 − 4 = 0. In the

latter case it√follows from equation (2) that x1 = 0, which in turn gives
that x2 = ± 2.
√
The equilibrium points are thus (0, 0) and (0, ± 2).
(b) The system equations on the set E simplifies to
ẋ1 = 4x21 x2
ẋ2 = −2x31 ,
since x21 + 2x22 − 4 = 0 on E. The set E is invariant since
d 2
( x + 2x22 − 4) = 2x1 ẋ1 + 4x2 ẋ2 = 2x31 4x2 + 4x2 (−2x31 ) = 0.
dt 1
for any x ∈ E. I.e. if x(T ) ∈ E for some T then x( t) ∈ E for all t ≥ T.

The motion on this invariant set is given by
ẋ1 = 4x2 · x21

ẋ2 = −2x1 · x21
(c) Use the squared distance to the set E as Lyapunov candidate
V ( x) = ( x21 + 2x22 − 4)2 .
Since V̇ = . . . = −4( x21 + 2x22 − 4)2 ( x1 f ( x1 ) + 2x2 f ( x2 )) we conclude

that V̇ < 0 everywhere except on E and x = 0. Since E is an ellipsoid
with the origin as its center, the state converges to E from any initial
condition except the origin (since distance to the E must decrease). I.e.
all trajectories except x " 0 tend to E.
(d) The set
√ E is NOT a limit cycle, as the two equilibrium points ( x1 , x2 ) =
(0, ± 2) belong to this set. Any trajectory moving along √ the invariant
set will
√ eventually end up in either ( x1 , x2 ) = (0, + 2) or ( x1 , x2 ) =
(0, − 2) .
Solution 3.8
Verify that V (0) = 0, V ( x) > 0 for x ,= 0 and V ( x) → ∞ for pp xpp → ∞. Now,
(a) We have
d
V ( x1 , x2 ) = 8x1 ẋ1 + 4x2 ẋ2 + 16x31 ẋ1 =
dt
= 8x1 x2 + 4x2 (−2x1 − 2x2 − 4x31 ) + 16x31 x2 =
= −8x22
Since V̇ ( x) ≤ 0, we conclude global stability.
71
(b) The Lyapunov function has V̇ ( x) = 0 for x2 = 0. For x2 = 0, we obtain
ẋ2 = −2x1 (2 + x21 ).
which implies that if x2 should remain zero, then x1 has also to be zero.
The invariance theorem from the lectures can now be used to conclude
global asymptotic stability of the origin.
Solution 3.9
(a) Introduce the state vector x = ( x1 , x2 )T = ( y, ẏ)T . The system dynamics
can now be written as
ẋ1 = x2
ẋ2 = −sat(2x1 + 3x2 )
Consider the Lyapunov function

Z 2x1 +3x2
1 1
V ( x1 , x2 ) = x22 + sat( z) dz
2 2 0
It is straight forward to show that V is positive definite. (Can x go to

infinity without V going to infinity?). Now,
d 1
V ( x1 , x2 ) = x2 ẋ2 + sat(2x1 + 3x2 )(2x1 + 3x2 )
dt 2
3
= − (sat(2x1 + 3x2 ))2
2
≤0
By Lyapunov theory, the system is globally stable. Further,
V̇ ( x) = 0 [ ẋ2 = sat(2x1 + 3x2 ) = 0

[ x1 ( t) = x2 ( t) = 0, ∀t
which implies global asymptotic stability.

(b) No. These conditions would imply global exponential stability. This can
not be achieved by the system, since in the saturated regions we have
ÿ = ±1.
(c) Try a slight modification of the procedure suggested in (a).

(d) No. So don’t work too long on this one.
Solution 3.10
72
The derivative of the suggested Lyapunov function is
V̇ ( x) = −2x2 max{0, x1 } max{0, x2 } ≥ 0
with equality for x1 ≤ 0, x2 ≤ 0 or both. Thus, the Lyapunov function

derivative is negative in the first quadrant and zero elsewhere in the plane.
When the Lyapunov function derivative is zero, we have
ẋ1 = x2
ẋ2 = − x1
This system has solutions
x1 ( t) = A cos( t) + B sin( t)
x2 ( t) = B cos( t) − A sin( t)
The trace of x1 , x2 is a circle, then if A and B are both nonzero, x1 ( t) > 0

and x2 ( t) > 0 for some t. This implies that the only solution of V̇ ( x) = 0 is
x( t) = 0 for all t. By LaSalle’s theorem, this system is globally asymptotically
stable.
Solution 3.11
(a)
1 0
5 6
P = 0.5
0 1
solves the Lyapunov equation with Q as the identity matrix.
Alternative:
V̇ = − x21 + x1 x2 − x1 x2 − x22 = −( x21 + x22 )
(b) We have
0
5 6
T T T
V̇ ( x) = x ( A P + P A) x + 2x P =
( x2 )
= − x21 − x22 + x2 ( x2 ) < 0
since the contribution from the x2 ( x2 )-term is non-positive under the

stated conditions.
(c) We have
V̇ ( x) = − x21 − x22 + x42
which is negative for x22 < 1. One might be tempted to consider the
whole strip
E = x : p x2 p < 1
) *
as a region of attraction. However, a couple of simulations show that

this is misleading, see Figure 7.5. The problem is that it is possible for
V to decrease even if p x2 p increases as long as p x1 p decreases sufficiently
fast. By taking a level set contained in E we guarantee that this does
73
Phase plane
2
1.5
0.5
x2
0
−0.5
−1
−1.5
−2
−2 −1.5 −1 −0.5 0 0.5 1 1.5 2
x1
Figure 7.5 Trajectories of the nonlinear system and level surfaces of V ( x) = 0.5xT x.
The region of attraction is the unit circle.
not happen. Since the level sets 0.5( x21 + x22 ) = γ are circles, we conclude
that the largest level set is
1
Ω = { x : V ( x) < }.
2
The unit circle is thus a guaranteed region of attraction.
Solution 3.12
(a) The origin is locally asymptotically stable, since the linearization
0 −1
5 6
d
x̃ = x̃
dt 1 −1
is asymptotically stable. A Lyapunov function for the system can be

found by solving the Lyapunov equation
AT P + P A = − I,
which has the unique solution
1.5 −0.5
5 6
P= (7.13)
−0.5 1
(You can solve Lyapunov equations in Matlab using the command

lyap(A.’,-eye(2));. Note the transpose on the A matrix, due to Mat-
lab’s definition of the command lyap.
(b) Since the Lyapunov function in (a) is positive for all x, we just have
to find the largest domain in which the derivative of the Lyapunov
function is negative. Introduce the polar coordinates (r, θ ) by
x1 = r cos θ
x2 = r sin θ
74
We get
V̇ (r, θ ) = −r2 + r4 cos2 θ sin θ (2 sin θ − cos θ ) ≤ −r2 + 0.861r4
which is negative for r2 < 1/0.861. Using this, together with λmin ( P ) ≥
0.69, we choose
0.69
c = 0.8 < = 0.801
0.861
The set { xp xT P x ≤ c} is an estimate of the region of attraction.
Solution 3.13
(a) For p2x1 + x2 p ≤ 1, we have
0 1
5 6
ẋ = x. (7.14)
−1 −1
The system matrix is ass. stable. Hence, the origin is locally asymptot-
ically stable.
(b) We have V ( x) > 0 in the first and third quadrant.
V̇ ( x) = ẋ1 x2 + x1 ẋ2 = x21 − x1 sat(2x1 + x2 ) + x22 .
Now, let us evaluate V̇ ( x) on the strip x1 x2 = c and suppose that c > 0

is chosen so large that the saturation is active, i.e., p2x1 + x2 p > 1.
By inspection of the dynamics, we see that for sufficiently large c,
no trajectories can escape from the first quadrant. We can therefore
use the following Lyapunov argument. Consider the Lyapunov function
derivative
c2
V̇ ( x) = x21 − x1 + .
x21
If c ≥ 1, V̇ ( x) is positive for all x1 ≥ 0. Hence, all trajectories starting

in the first quadrant to the right of the curve x1 x2 = c cannot cross the
curve. Consequently, they cannot reach the origin.
(c) It follows from (b) that the origin is not globally asymptotically stable.
75
Solution 3.14
(a) Use V as a Lyapunov function candidate and let u be generated by the
nonlinear state feedback
V
u = −( ψ ( x))
x
(b) Intentionally left blank.
Solution 3.5
Use convexity wrt K .
Solution 3.16
f
(a) Integration of the equality d
dσ f (σ x) = x (σ x) · x gives the equation
Z1
f
f ( x) = (σ x) · x dσ .
0 x
We get
Z1 Z1 5 6T
T T f T T f
x P f ( x) + f ( x) P x = x P (σ x) xdσ + x (σ x) dσ P x
0 x 0 x
Z1 I 5 6T J
T f f
=x P (σ x) + (σ x) P dσ x ≤ − xT x
0 x x
(b) Since P is positive definite, V ( x) is clearly positive semidefinite. To

show that it is positive definite, we need to show that f ( x) = 0 only
when x = 0. But the inequality proved in (a) shows that if f (p) = 0
then
0 ≤ −pT p.
(c) Suppose that f is bounded, i.e. that q f ( x)q ≤ c for all x. Then
q xT P f + f T P xq ≤ 2cq P qq xq.
But this contradicts the inequality in (a) as q xq → ∞.

(d) We have shown that V is positive definite and radially unbounded.
Moreover
5 6T C 3 4T D
f f f f
V̇ = ẋ T T
P f +f P ẋ = f T
P ( x) + ( x) P f ≤ −q f ( x)q2 .
x x x x
Hence V̇ ( x) < 0 for all x ,= 0. Thus, the origin is globally asymptotically

stable.
76
Solution 3.17
Assume the linearization A = xf of f is asymptotically stable. Then the
equation
P A + AT P = − I,
R∞ T
has a solution P > 0. (To prove that P = 0
e A s e As ds > 0 is such a solution
integrate both sides of
d AT s As T T
e e = AT e A s e As + e A s e As A
ds
from 0 to ∞.) All conditions of Krasovskii’s method are then satisfied and we
conclude that the nonlinear system is asymptotically stable. The instability
result is harder.
Solution 3.18
The system is given by
ẋ1 = x2 =: f1
ẋ2 = − x2 + K ( e) = − x2 + K (− x1 ) =: f2 .
Following the hint we put V = f T ( x) P f ( x). Direct calculations give

 
 −6p 12 K x21 p 11 − p 12 − 3 K p 22 x21 
V̇ = f 
T
 
 f.
p 11 − p 12 − 3 K p 22 x21 2(p 12 − p 22 ).
With  
 1 1
P=
 

1 2
we get V̇ ≤ 0 if 3 K x21 < 1. Hence the system is locally stable. Actually one
gets V̇ < 0 if 3 K x21 < 1 unless x1 = 0. The invariant set is x1 = x2 = 0. From
LaSalle’s theorem the origin is hence also locally asymptotically stable.
77
Solution 4.1
See the Figure 7.6.
1.5 1.5
1 1
0.5 0.5
0 0
−0.5 −0.5
−1 −1
−1.5 −1.5
−1 0 1 −1 0 1
1.5 1.5
1 1
0.5 0.5
0 0
−0.5 −0.5
−1 −1
−1.5 −1.5
−1 0 1 −1 0 1
Figure 7.6 The balls B(0, 1) in Exercise 4.1
Solution 4.2
(a) What are the restrictions that we must impose on the nonlinearities so
that we can apply the various stability theorems?
The Nyquist Criterion ψ ( y) must be a linear function of y, i.e., ψ ( y) =
k 1 y for some constant k 1 .
The Circle Criterion ψ ( y) must be contained in some sector [ k 1 , k 2 ].
Small Gain Theorem ψ ( y) should be contained in a symmetric sector
[− k 2 , k 2 ]. The gain of the nonlinearity is then less or equal to k 2 .
The Passivity Theorem states that one of the systems must be strictly
passive and the other one passive. Here we consider the case where
ψ is strictly passive. Let y = ψ (u). According to the definition in
the lecture notes a system is strictly passive if
(u, y)T ≥ ε (quq2T + q yq2T )
for all u and T > 0 and some ε > 0. This requires ψ (0) = 0, and
since ψ is static the following must hold for any input u:
y(u)u ≥ ε (u2 + y(u)2 ) ∀t ≥ 0
The last inequality can, through dividing by y2 and carrying out

completion of squares, also be written as
u 1 1
( − )2 ≤ 2 − 1
y 2ε 4ε
78
We first note that the inequality can only hold of ε ≤ 1/2, and
in the following we assume that ε ≤ 1/2. Let x = u/ y. Then the
inequality reads
1 1 − 4ε 2
( x − )2 ≤
2ε 4ε 2
And we note that it holds for
1
x= [ y = 2εu
2ε
as then the LHS is zero. If we pick
u
x = 2ε [ y =
2ε
we get for the LHS

42 42
1 4ε 2 − 1
3 3
x− =
2ε 2ε
which is smaller than the RHS as
(4ε 2 − 1)2 ≤ 1 − 4ε 2
for ε < 1/2. Finally, the inequality holds for all
1
2ε ≤ x ≤
2ε
as the derivative of the LHS is given by
1 1
x− ≤0 ∀x ≤ .
2ε 2ε
Thus a necessary condition is that y(u) satisfy
1
2εu ≤ y(u) ≤ u
2ε
which gives that φ must belong to the sector [ε, ε1 ] for some small
ε>0
These conditions are illustrated in Figure 7.7.
(b) If the above restrictions hold, we get the following conditions on the
Nyquist curve
The Nyquist Criterion The Nyquist curve should not encircle the
point −1/ k 1 .
The Circle Criterion If 0 ≤ k 1 ≤ k 2 , the Nyquist curve should
neither encircle nor intersect the disc defined by −1/ k 2 , −1/ k 1 . If
k 1 < 0 < k 2 G should stay inside the disc.
Small Gain Theorem The Nyquist curve has to be contained in a disc
centered at the origin, with radius 1/ k 2 .
79
Nyquist criterion Circle Criterion

1 1
k
2
0.5 k 0.5
1
k
1
ψ(y)
ψ(y)
0 0
−0.5 −0.5
−1 −1
−1 −0.5 0 0.5 1 −1 −0.5 0 0.5 1
y y
Small Gain Theorem Passivity theorem

1 1
k
2
0.5 0.5
ψ(y)
ψ(y)
0 0
−k
−0.5 2 −0.5
−1 −1
−1 −0.5 0 0.5 1 −1 −0.5 0 0.5 1
y y
Figure 7.7 Sector conditions on memoryless nonlinearity.
Nyquist criterion Circle Criterion

1 1
0.5 0.5
ψ(y)
ψ(y)
0 0
−0.5 −0.5
−1 −1
−1 −0.5 0 0.5 1 −1 −0.5 0 0.5 1
y y
Small Gain Theorem Passivity theorem

1 1
0.5 0.5
ψ(y)
ψ(y)
0 0
−0.5 −0.5
−1 −1
−1 −0.5 0 0.5 1 −1 −0.5 0 0.5 1
y y
Figure 7.8 Conditions on Nyquist curve matched to sector conditions on memoryless

nonlinearity.
The Passivity Theorem Since we assumed that ψ is strictly passive,

G needs to be passive. Thus the Nyquist curve has to stay in the
right half-plane, Re(G(iω )) ≥ 0.
These conditions are illustrated in Figure 7.8.
(c) The Passivity theorem and Small gain theorem can handle dynamic
nonlinearities.
80
Solution 4.3
(a) The systems belong to the sectors [0, 1], [0, ∞] and [−1, ∞] respectively.
(b) Only the saturation nonlinearity (the leftmost nonlinearity) has finite
gain, which is equal to one. The other two nonlinearities have infinite
gain.
(c) The nonlinearity is passive if uy ≥ 0. That is if and only if the curve is
contained in the first and third quadrants. The saturation and the sign
nonlinearity are passive. The rightmost nonlinearity is not passive.
Solution 4.4
Since the linear part of the system is Hurwitz, we are free to use all versions
of the circle criterion.
(a) In order to guarantee stability of a nonlinearity belonging to a sym-
metric sector [−α, α], the Nyquist curve has to stay strictly inside a
disk centered at the origin with radius 1/α. We may, for instance, take
α = 0.25 − ε for some small ε > 0.
(b) The Nyquist curve lies inside the disk D(−1.35, 4.35). Thus, stabil-
ity can be guaranteed for all nonlinearities in the sector −0.23, 0.74.
(NOTE: The disk D( x1 , x2 ) is defined as the disk with diameter p x1 − x2 p
which crosses the real axis in x1 and x2 .)
(c) We must find β such that the Nyquist plot lies outside of a half-plane
Re(G(iω )) < −1/β . A rough estimate from the plot is β = 1.1.
Solution 4.5
The open loop system has one unstable pole, and we are restricted to apply
the first or fourth version of the circle criterion. In this example, we can place
a disk with center in −3 and with radius 0.75, and apply the first version of
the Nyquist criterion to conclude stability for all nonlinearities in the sector
[0.27, 0.44].
Solution 4.6
(a) The circle with k 1 = −2, k 2 = 7 does not intersect the Nyquist curve
(see Figure 7.9). Hence the sector (−2, 7) suffices. As always there are
many other circles that can be used (The lower limit can be traded
against the upper limit).
(b) The Nyquist diagram is a circle with midpoint in −0.5 and radius
0.5, see Figure 4.5. Since the open system has two unstable poles the
Nyquist curve should encircle the disc twice. Choosing the circle that
passes through −1/ k 1 = −1 + ε and −1/ k 2 = −ε, we conclude by the
1
Bode-diagram, that the loop is stable for the sector [ , 1/ε ],
1−ε
(One might think that the Bode-diagram only indicates one encir-
clement. However, the Bode-diagram is only for positive frequencies,
and will be mirrored for negative frequencies, yielding two encirclements.)
81
Nyquist Diagrams
0.5
0.4
0.3
0.2
0.1
Imaginary Axis
0
−0.1
−0.2
−0.3
−0.4
−0.5
−0.5 −0.4 −0.3 −0.2 −0.1 0 0.1 0.2 0.3 0.4 0.5
Real Axis
Figure 7.9 The Nyquist curves for the system in Exercise 4.6a, and the circle
corresponding to k1 = −2, k2 = 7.
82
Solution 4.7
(a) Introduce y = C x and u = δ y = ψ ( y), then
ẋ = Ax + Bu
y = Cx
Y (s) = C X (s) = C (s I − A)−1 BU (s) = G(s)U (s)
(b) ψ satisfies:
Z∞ Z∞
qψ ( y)q22 = pδ ( t) y( t)p2 dt = pδ ( t)p2 p y( t)p2 dt
0 0
Z∞
2
≤ sup pδ ( t)p p y( t)p dt = sup pδ ( t)p2 q yq22 ≤ q yq22
2
t 0 t
Thus γ (ψ ) ≤ 1. The gain of the linear system is γ (G). Then, according

to the Small Gain Theorem the feedback connection is BIBO-stable if
γ (G) < 1. Since the gain of a linear system is given by
γ (G) = sup pG(iω )p < 1 (7.15)

ω ∈(0,∞)
the desired results follow directly.

(c) Only the leftmost Nyquist curve shows a system with gain greater
than one. Thus, systems corresponding to the middle and the rightmost
Nyquist curve are guaranteed to give a BIBO stable feedback loop.
(d) Follows from the definition of gain for linear time invariant MIMO
systems.
83
Solution 4.8
(a) >> A=[1 10; 0 1];svd(A)
ans =
10.0990
0.0990
(b)
3 4
q ABxq q ABxq q Bxq
σ1 ( AB) = sup = sup ·
x q xq x q Bxq qx
3 4
q Ayq q Bxq
sup · sup = σ1 ( A)σ1 ( B)
y q yq x q xq
Solution 4.9
The proof follows directly from the definition of passivity, since, according to
the definition of a storage function
ZT
〈u, y〉T = uT y dt
0
ZT
≥ V̇ ( x) dt = V ( x(T )) − V ( x(0)) = V ( x(T ))
0
which is non-negative since V ( x(T )) ≥ 0. Passivity follows.
84
Solution 4.10
The linear system G(s) corresponds to
ẋ = Ax + Bu, y = BT P x, x(0) = 0.
Let V = xT P x. Then
V̇ = ẋT P x + xT P ẋ
= xT ( AT P + P A) x + 2xT P Bu = − xT x + 2yT u ≤ 2yT u
Integrate and use the fact that V (0) = 0, then

ZT
yT udt ≥ V (T ) − V (0) ≥ 0,
0
which proves passivity.
Solution 4.11
Write the system in state-space form:
ẋ = −2x + sat( x) + u, x(0) = 0

y=x
We try using the quadratic storage function V ( x) = x2 /2, which is typically

a good start when looking for a function to show passivity. First note that
d
V = x(−2x + sat( x) + u)
dt
= yu − 2x2 + xsat( x) ≤ xu − x2
as
x2 ≥ xsat( x) ≥ 0.
(a)
d
V ≤ yu − x2 ≤ yu
dt
Hence, the system is passive.

(b) As we have that dt
d
V ≤ yu − x2 where x2 is a positive definite function
(which is zero only if x = 0), it follows that the system also is strictly
passive.
(c) You can solve this problem in many ways. Here we will give two alter-
natives: One using passivity and one based on the circle criterion.
Alternative 1:(Passivity) The controller looks very much as the strictly
passive system from (a) and (b), and we therefore introduce a new
variable x2 = z − θ :
85
ẋ2 = ż − θ˙ = 2(θ − z) − sat(θ − z) − ω

= −2x2 + sat( x2 ) − ω
η = z − 2θ = x2 − θ
This will be the block-scheme according to Figure 7.10. We see that the
strictly passive system Σ c with input ω and output x2 will be feedback
connected to another subsystem which consists of the DC-motor with
a local feedback with −θ ( coming from one term of η). The transfer
function of this subsystem will be
1
s+1 s
=
1 + 1s s+1 1 s2 +s+1
which is a passive system. We thus have a passive system in feed-

back with a strictly passive system and therefore,see lecture notes, the
closed-loop system will be asymptotically stable, which means that both
ω and θ approach 0 as t → ∞.
η 1 ω 1 θ
+ s+1 s
−1
x2
Σc
Figure 7.10
Alternative 2:(Circle criterion) With the obvious state vector x =

(θ , ω, z)′ , we rewrite the system in the feedback connection form
0 1 0 0
   
ẋ = Ax − Bψ ( y) =  −2 −1 1  x −  0  sat([ 1 0 −1 ] x)
2 0 −2 1
The Nyquist curve of the linear system is illustrated in Figure 7.11.

Since the Nyquist curve does not intersect the half plane Re(G(iω )) <
−1/2, we conclude stability for all ψ in the sector [0, 2]. Due to that
the saturation element lies in the sector [0, 1], we conclude asymptotic
stability of the closed loop.
86
Nyquist Diagrams
0.8
0.6
0.4
0.2
Imaginary Axis
0
−0.2
−0.4
−0.6
−0.8
−1
−1.5 −1 −0.5 0 0.5
Real Axis
Figure 7.11 Nyquist curve of linear subsystem.
Solution 4.12
(a) We have
ZT
〈 y, u〉 = y( t)u( t) dt =
0
ZT
= {u( t)uc ( t)}{ H (u( t)uc ( t))}dt =
0
ZT
= w( t) H (w( t)) dt = 〈w, H (w)〉
0
where w = uc u. Since H is passive, the result follows.

(b) We will only consider the case where θ 0 = 0. The case θ 0 is a little
tricker, and the discussion about this case is postponed to the course in
Adaptive Control.
If θ 0 = 0, the system equations read
e( t) = G(p)θ uc ( t)
θ˙ ( t) = −γ uc ( t) e( t)
In light of exercise (a), we can identify the second equation modeling

the signal w( t) = uc ( t) e( t) sent into an integrator with gain γ and
postmultiplied by uc (See the lecture slides for an illustration). This
system is passive, and interconnected in a negative feedback loop with
the strictly passive system G. Stability now follows from the passivity
theorem.
Solution 4.13
No solution yet.
87
Solution 5.1
Use the interpretation of describing function as "equivalent gain" and analyse
the gains of each non-linearity sectionally. We have 1-b, 2-c, 3-a, 4-d.
Solution 5.2
Denote the nonlinearity by f . For memoryless, static nonlinearities, the de-
scribing function does not depend on ω, and the describing reduces to
b1 ( A) + ia 1 ( A)
N ( A) =
A
where a 1 and b1 can be computed as

Z 2π
1
a1 = f ( A sin(φ )) cos(φ ) dφ (7.16)
π 0
Z 2π
1
b1 = f ( A sin(φ )) sin(φ ) dφ. (7.17)
π 0
(a) First, we notice that the saturation is an odd function, which implies
that a 1 = 0. In order to simplify the computations of b1 , we set H = 1
and note that the saturation can be described as
A/ D sin(φ ) 0 ≤ φ ≤ φl
I
f ( A sin(φ )) =
1 φ l < φ < π /2
Here, φ l = arcsin( D/ A) denotes the value of φ where so that f satu-

rates. Now,
Z 2π
1
b1 = u(φ ) sin(φ ) dφ =
π 0
Z π /2
4
= u(φ ) sin(φ ) dφ =
π
A0Z Z π /2 B
4 φl
= A/ D sin2 (φ ) dφ + sin(φ ) dφ =
π 0 φl
AZ
φl Z π /2 B
4
= A/(2 D)(1 − cos(2φ )) dφ + sin(φ ) dφ =
π 0 φl
4
= ( A/(2 D)(φ l − sin(φ l ) cos(φ l )) + cos(φ l )) =
π 3
2A
4
D
= φl + cos(φ l )
Dπ A
Thus, the describing function for the normalized saturation is
2 D
N ( A) = (φ l + cos(φ l ))
Dπ A
88
Now, using the calculation rule Nα f ( A) = α N f ( A), we find that for

the saturation under consideration we have
2H D
N ( A) = (φ l + cos(φ l ))
Dπ A
(b) We see that the nonlinearity is a superposition of a linear function

H
( e) = e
D
and the nonlinearity − f ( e) with f ( e) as in (a). Using the fact that a
linear function ( e) = ke has describing function N ( A) = k, and the
superposition rule N f + ( A) = N f ( A) + N ( A), we find
2
3 ; <4
H D
N ( A) = 1− φl + cos(φ l )
D π A
(c) Noting that this nonlinearity can be written as the sum of the two
nonlinearities in (a) and (b), we arrive at the describing function
2(α − β )
3 4
D
N ( A) = φl + cos(φ l ) + β .
π A
Solution 5.3
Let the input to the relay be
u( t) = A sin(ω t) = A sin(φ )
The output of the relay is then



 −H 0 < φ < φ0

y(φ ) = H φ0 < φ < π + φ0



−H π + φ 0 < φ < 2π
where φ 0 = arcsin( D/ A). We get

Z 2π
1
a1 = y(φ ) cos(φ ) dφ
π 0
Z π +φ 0 Z 2π
1 1 1
Zφ0
= (− H ) cos(φ )dφ + H cos(φ ) dφ + (− H ) cos(φ )dφ
π 0 π φ0 π π +φ 0
4H
=− sin(φ 0 )
π
and
Z 2π
1
b1 = y(φ ) sin(φ ) dφ
π 0
Z π +φ 0 Z 2π
1 1 1
Zφ0
= (− H ) sin(φ )dφ + H sin(φ ) dφ + (− H ) sin(φ )dφ
π 0 π φ0 π π +φ 0
4H
= cos(φ 0 )
π
89
We obtain
4H
N ( A) = (cos(φ 0 ) − i sin(φ 0 ))
πA
p
The identity cos( z) = 1 − sin2 ( z) gives the desired result.
Solution 5.4
Follows from the integration rule
1
Z
f ( ax) d x = F ( ax)
a
where F ( x) =
R
f ( x) d x.
Solution 5.5
We have
φ ( x) φ ( a)
< , x < a.
x a
and thus
π
2
Z
Φ( a) = φ ( a sin(θ )) sin(θ ) dθ
aπ 0
Zπ
2 φ ( a)
< a sin(θ ) sin(θ ) dθ
aπ 0 a
2
Zπ
= φ ( a) sin2 (θ ) dθ = φ ( a)/ a
aπ 0
Solution 5.6
The describing function is
N ( A) = k 1 + 3 A2 k 3 /4
Note, however, that the output y(T ) of the nonlinearity for the input e( t) =
A sin(φ ) is
y( t) = A2 k 2 /2 + ( k 1 A + 3 A3 k 2 /4) sin(φ )
− A2 k 2 /2 · cos(2φ ) − A3 k 3 /4 · sin(3φ )
We conclude that the term k 2 x22 does not influence N ( A). Still, we can not
just apply the describing function method, since there is a bias term. If the
linear system has integral action, the presence of a constant offset on the
input will have a very big influence after some time.
90
Solution 5.7
(a) When the saturation works in the linear range, we have the closed loop
dynamics
−5s
G(s) =
s2 + (1 − 5)s + 25
which is unstable. Thus, the state can not remain small. In saturation,
on the other hand, the nonlinearity generates a constant(“step”) input
to the system. The final value theorem then gives
−5s
lim y( t) = lim =0
t→∞ s→0 s2 + s + 25
The observation that y( t) → 0 contradicts the assumption that the

nonlinearity remains saturated.
(b) We should investigate intersection of the Nyquist curve and −1/ N ( A).
From Exercise 5.2 we have that

A A<1
N ( A) =
 2 φ l + 1 cos(φ l ) A≥1
! "
π A
with φ l = arcsin(1/ A). We know that limit cycles can only occur at sat-
uration, so we only consider A ≥ 1. Then N ( A) ∈ (0, 1] and −1/ N ( A)
lies in the interval (−∞, −1].
The frequency response of the system is
−i5ω −5ω 2 5ω (ω 2 − 25)

G(iω ) = = + i (7.18)
25 − ω 2 + iω (25 − ω 2 )2 + ω 2 (25 − ω 2 )2 + ω 2
which intersects the negative real axis for ω ′ = 5 rad/s. The value of
G(iω ′ ) = −5. Thus, there will be an intersection. The frequency of the
oscillation is estimated to 5 rad/s, the amplitude is given by
1
− = G(iω ′ ) = −5 [ N ( A) = 0.2
N ( A)
From Figure 5.6 we see that A = 6.

(c) From (7.18) we see that the Re(G) ≤ 0 for all ω ∈ R, and that Im(G) <
0 for ω ∈ [0, 5), and Im(G) ≥ 0 for ω ≥ 5. The Nyquist curve of the
system is shown in Figure 7.12. The function −1/ N ( A) is situated
on the negative real axis between −∞ and −1 (again, for A > 1).
−1 is maked by a small line. The Nyquist curve encircles the points
Re(G(iω )) > −5, indicating increased oscillation amplitude. The points
to the left of the intersection are not encircled, indicating stability and a
decaying oscillation amplitude. We can thus expect a stable limit cycle.
91
Nyquist Diagrams
Imaginary Axis
0
−1
−2
−3
−4
−8 −7 −6 −5 −4 −3 −2 −1 0 1
Real Axis
Figure 7.12 Nyquist curve and −1/ N ( A) for oscillator example.
Solution 5.8
(a) Introduce θ 0 = arcsin( a/ A) and proceed similarly to the saturation
nonlinearity.
(b) The describing function has maximum for
√
A∗ = 2a
which gives
2
N ( A∗ ) =
πa
√
The Nyquist curve
√ crosses the negative real axis for ω = 2, for which
the gain is G(i 2) = −2/3. Thus, we should expect no oscillations if
4
a> .
3π
Solution 5.9
(a) The describing function for a relay with amplitude D is given by
4D
N ( A) =
πA
−1/ N ( A) lies on the negative real axis. If the Nyquist curve intersects
the negative real axis, the describing function methods will predict a
sustained oscillation
4D
− pG(iω u )p = −1
πA
Thus, given the amplitude A of the oscillation, we estimate the ultimate
gain as
4D
K u = 1/pG(iω u )p =
πA
The ultimate period is the period time of the oscillations
Tu = 2π /ω
92
(b) From the simulation, we estimate the amplitude A = 0.6 which gives
K u ( 2.12. The ultimate period can be estimated directly from the plot
to be Tu ( 2. Note that the estimates have good correspondence with
the analytical results (which require a full process model)
Solution 5.10
No solution yet.
Solution 5.11
No solution yet.
Solution 5.12
No solution yet.
Solution 5.13
No solution yet.
93
Solution 6.1
We would like to write the system equations as
v = G(s)(−u)
u = φ (v)
where φ (·) denotes the saturation. Block diagram manipulations give

3 4
AR BS
v=u− + u
Aw A Aw A
3 4
AR + BS
= − 1 (−u) = G(s)(−u)
A Aw
Since the saturation element belongs to the sector [0, 1], we invoke the circle
criterion and conclude stability if the Nyquist curve of G(iω ) does not enter
the half plane Re(G(iω )) < −1. This gives the desired condition.
Solution 6.2
The model is given by
dz pvp
=v− z (7.19)
dt (v)
dz
F = σ0 z + σ1 (v) + Fv v (7.20)
dt
(a) For constant velocity v ,= 0, the stationary point z∗ is
(v)
z∗ = v = (v)sign(v).
pvp
z∗ is asymptotically stable since
d ( z − z∗ ) pvp
=− ( z − z∗ ),
dt (v)
where (v) ≥ 0.
(b) For any constant velocity, v ,= 0, (7.19) converges to z = (v)sign(v),
and F therefore converges to
F = σ0 (v)sign(v) + Fv v
(c) Following the hint, we consider V ( z) = z2 . Along trajectories of the

model, we have
pvp
V̇ = 2z(v − z)
(v)
p zp
≤ 2p zppvp(1 − )
(v)
94
which is non-positive if p zp/(v) ≥ 1. Since 0 ≤ (v) ≤ a, we see that

for
p z( t)p ≥ a
we have V̇ ≤ 0. We conclude that
Ω = { zp zT z < a2 }
is invariant. In other words, all trajectories that start within Ω remain

there for all future times. The set Ω provides a bound on the state z( t).
(d) Consider the storage function V ( z) = z2 /2. We have
dz pvp 2 dz
zv = z + z ≥z = V̇ ( t)
dt (v) dt
and passivity follows from the discussion in (+).

(e) We have
pvp
Fv = Fv v2 + (σ1 ż + σ0 z)( ż + z) (7.21)
(v)
pvp pvp
≥ σ1 ż2 + σ0 z2 + ( σ1 + σ0 ) z ż (7.22)
(v) (v)
In the expression above we recognize the term σ0 z ż as the time deriva-

tive of the storage function V ( z) = σ0 z2 /2. Next, we separate out the
storage function derivative and make a completion of squares to esti-
mate the additional terms
pvp 2 pvp
Fv ≥ σ0 z ż + σ1 ż2 + σ0 z + σ1 z ż (7.23)
(v) (v)
3 42 A 3 42 B
pvp pvp pvp
= V̇ + σ1 ż + z + σ0 − σ1 z2 (7.24)
2 (v) (v) 2 (v)
Since the second term is non-negative, we have
Fv ≥ V̇
and thus passivity if
pvp
σ0 − σ1 >0
4 (v)
This concludes the proof.
Solution 6.3
(a) The describing function for a relay has been derived on Lecture 6 to be
4 F0
N ( A) =
πA
95
(b) Using the superposition property of describing functions for static non-
linearities N f + = N f + N , and the fact that for a scalar gain y = ku
the describing function is N ( A) = k, we obtain
4 F0
N ( A) = Fv +
πA
(c) Stiction is a point-wise phenomenon (occurring for v = 0) with finite

amplitude, and has no influence on the integral calculations involved in
the describing function computation. The describing function is there-
fore the same as in (b).
Solution 6.4
The process is given by
ẋ = v
v̇ = − F + u
The velocity is observed through
b = zv + K v x
v
b + u − K v vb
żv = − F
b denotes the estimated friction force, estimated by the observer

where F
b = ( z F + K F pb
F vp)sign(b
v)
b)sign(b
ż F = − K F (u − F v)
Defining the observer errors
e v = v − vb
b
eF = F − F
we obtain the observer dynamics
b + u − K v vb) − K v v
ė v = v̇ − vḃ = v̇ − żv − K v ẋ = − F + u − (− F
b − K v (v − vb) = − e F − K v e v
= −F + F
1 2
ḃ = Ḟ − ż F sign(b
ė F = Ḟ − F b)
v) − K F vḃ = Ḟ − − K F (u − F
! "
− K F − F̂ + u − K v vb + K v v = Ḟ − K F K v (v − vb) = Ḟ − K F K v e v
The term Ḟ is zero (except at zero velocity where it is not well defined).
Putting Ḟ = 0, we obtain
− Kv −1
5 6 5 65 6
ė v ev
= (7.25)
ė F − Kv K F 0 eF
with the characteristic equation
λ(s) = s2 + K v s − K v K F
96
We conclude that the error dynamics are locally asymptotically stable if
K v > 0,
− Kv K F > 0
which implies the desired conditions.
97
D
− D/2
D/2
−D
Figure 7.13 The function in 6.5.
Solution 6.5
(a) The gain margin for the system is 1.33 > 1.27, thus there should
be no limit cycle since the gain margin exceeds that required for the
worst-case scenario with quantization.
(b) We have already (lecture) seen that the describing function for the
function in Figure 7.13 is given by
0 A < D/2
I
N D ( A) = q
4D
πA 1 − ( 2DA )2 A > D/2
Superposition ( N f1 + f2 = N f1 + N f2 ) gives
NQ = N D + N3 D + N5 D + . . . + N2i+1
which gives the stated describing function NQ ( A) for the quantizer.
Solution 6.6
We have
ZT ZT
〈u, y〉T = uy dt = usat(u) dt ≥ 0
0 0
We conclude passivity from to the definition given in the lecture slides.
Solution 6.7
No solution yet.
Solution 6.8
Assume without loss of generality that 0 < u0 < D/2. The input to the
quantizer is u0 + d ( t) where d ( t) is the dither signal. The output y from the
quantizer is
I
0 u0 + d ( t) < D/2
y( t) = Q(u0 + d ( t)) =
D u0 + d ( t) > D/2
98
It is easy to see that y = D during a time interval of length uD0 T, where T is

the time period of the dither. The average value of y becomes
1 u0
y0 = T · D = u0 .
T D
Hence the dither signal gives increased accuracy, at least if the signal y can
be treated as constant compared to the frequency of the dither signal. The
method does not work for high-frequency signals y.
Solution 6.9
No solution yet.
Solution 6.10
No solution yet.
99
Solution 7.1
Let the output of the nonlinearity be u, so that u = f (v).
(a) We have
u = v2 , v≥0
which implies that

√
v= u, u≥0
(b) The piecewise linear characteristic

I
k 1 v, pvp ≤ d
u=
sign(v)( k 1 − k 2 ) d + k 2 v pvp > d
gives the inverse
I
u/ k 1 , pup ≤ k 1 d
v=
(u − sign(u)( k 1 − k 2 )d )/ k 2 pup > k 1 d
Consider a (unit) dead-zone with slope k 1 = ε in the interval pvp ≤ d,
and slope k 2 = 1 otherwise. We obtain the inverse
I
u/ε, pup ≤ ε d
v=
u + sign(u)(1 − ε ) d, pup > ε d
The dead-zone nonlinearity and its inverse are shown in Figure 7.14.
Dead zone Dead zone inverse

2 2
1 1
v=f−1(u)
u=f(v)
0 0
−1 −1
−2 −2
−2 0 2 −1 0 1
v u
Figure 7.14 Deadzone and its inverse.
(c) We need xin to jump ±2 D when xout changes sign. xout will change sign
if u goes from increasing to decreasing. Thus the following inverse will
work: 
b if u( t) > u( t−)
u + D

xin ( t) = u− D b if u( t) < u( t−)


xin ( t−) otherwise
See Figure 7.15 for an illustration.
100
xin
b
D
u
b
−D
Figure 7.15 Backlash Inverse
Solution 7.2
(a) We notice that all state equation but the last one are linear. The last
state equation reads
ẋn = f ( x) + ( x)u
If we assume that ( x) ,= 0 for all x, we can apply the control
1
u = h( x, v) = (− f ( x) + Lx + v)
( x)
renders the last state equation linear
ẋn = Lx + v
The response from v to x is linear, and the closed loop dynamics is

given by
0 1 0 0 0
 0
0 0 1 0 ... 0
ẋ =  ..  x +  ..  v
   
0 0 0 . 0  .
l1 l2 l3 ... ln 1
(You may recognize this as the controller form from the basic control
course). For the control to be well defined, we must require that ( x) ,=
0 for all x.
(b) The above procedure suggest the control
1
u= (− a sin( x1 ) + l 1 x1 + l 2 x2 + v)
b cos( x1 )
which results in the closed loop system
0 1 0
5 6 5 6
ẋ = x+ v
l1 l2 1
101
The system matrix has a double eigenvalue in s = −1 if we let
l 1 = −1, l 2 = −2
The control law is well defined for x1 ,= π /2. This corresponds to the
pendulum being horizontal. For x1 = π /2, u has no influence on the
system. Notice how the control “blows up” nearby this singularity. Extra.
You may want to verify by simulations the behaviour of the modified
control
u = sat( h( x, v))
for different values of the saturation level.

(c) The above procedure suggest the control
u = − x2 − x + v
Letting v = 0, we apply the control to the perturbed system
ẋ = (1 + ε ) x2 − x2 − x = ε x2 − x
and note that for x > 1/ε, we have ẋ > 0, which implies that the
trajectories tend to infinity. Thus, global cancellation is non-robust in
the sense that it may require a very precise mathematical model.
Solution 7.3
(a) The sliding set in a sliding mode design is invariant, i.e., if x( t s ) belongs
to the sliding surface σ ( x) = 0, at time t s , then it belongs to the set
σ ( x) = 0 for all future times t ≥ t s . Thus, it must hold that
σ ( x) = σ̇ ( x) = 0
which yields the dynamics on the sliding set.

(i) We have
σ̇ ( x) = 2 ẋ1 − ẋ2 = 2 ẋ1 − x1 = 0
The third equality implies that x1 ( t) → ±∞ on the sliding set.

Thus, forcing this surface to be a sliding mode would give unstable
solutions.
(ii) Similarly as above
σ̇ ( x) = ẋ1 + 2 ẋ2 = ẋ1 + 2x1 = 0
Thus, the equivalent dynamics along this surface satisfy ẋ1 =

−2x1 and is hence asymptotically stable.
(iii) We have
σ̇ ( x) = ẋ1 = 0
The dynamics on this sliding set would thus be stable, but not
102
(b) According to the lecture slides, the sliding mode control law is
pT Ax µ
u=− T
− T sign(σ ( x))
p B p B
Where the sliding set is given by
σ ( x) = pT x = 0
Thus, in this example we have pT = [ 1 1 ] and
u = −( x1 + x2 ) − µsign( x1 + x2 )
(c) According to the robustness result of the sliding mode controller pre-
sented on the lecture, the above controller will force the system toward
the sliding mode if µ is chosen large enough, and if sign(pT B̂) =
sign(pT B), which implies sign(b̂) = sign(b). Since the nominal design
has b̂ = 1, we must have
b>0 (7.26)
It remains to check that the dynamics of the sliding mode remains

stable. (Otherwise, we could have a situation where the controller forces
the state onto the sliding mode, but where the sliding mode dynamics
are unstable. The state would then tend toward infinity along the
sliding mode.) In this case, we can verify that on the sliding mode, we
have σ̇ ( x) = 0 for all values of the parameter a.
Solution 7.4
(a) Straightforward manipulations give
K 1
G(s) = e−sL = e−sVd /q
sT + 1 sVm /q + 1
(b) The step response gives parameters a = 0.9, L = 1. Using the results
from (a) and a = K L/T we obtain
a = Vd / Vm
L = Vd /q
Since the experiment was performed for q = 1, we see that L = Vd .

Now, a gain scheduled PI controller can be constructed using Ziegler-
Nichols recommendations as
K p = 0.9/ a = 1
Ti = 3L = 3/q
Here we see that K p remains constant wheras Ti changes with the flow
q.
103
Solution 7.5
(a) The pendulum energy is given by
Jp 2
E ( x) = m l(1 − cos( x1 )) + x
2 2
If the energy of the pendulum hanging downwards is taken to be E (0) =

0, the energy for x1 = π, x2 = 0 is E 0 = 2m l.
(b) The time derivative of the Lyapunov function candidate reads
d
V̇ ( x) = 2( E ( x) − E 0 ) E ( x) =
dt
= 2( E ( x) − E 0 )(m l sin( x1 ) ẋ1 + Jp x2 ẋ2 ) =
= 2( E ( x) − E 0 )(−mlx2 cos( x1 )u)
Applying the suggested control, we obtain
V̇ ( x) = −2kml( E ( x) − E 0 )2 x2 cos( x1 )sign( x2 cos( x1 )) ≤ 0
if k > 0, with equality attained for E ( x) = E 0 , or x2 = 0 or x1 = π /2.

The only unwanted invariant manifold is x1 = x2 = 0.
(c) The phase portrait of the closed loop system is shown in Figure 7.16.
We notice how the state is driven to the set E ( x) = E 0 , and that this
set contains no stable equilibrium points. Note that the velocity ap-
proaches zero as the pendulum approaches the upright position. Since
the equilibrium point is unstable, and the control for this state is zero,
the pendulum does not remain in the upright position.
Extra. Feel free to design a stabilizing controller for the upright position
(using, for example the results from Exercise 7.2). In particular, how
should you switch between the two control strategies to make the system
stable? (Some Lyapunov theory will help you on this one)
Phase plane
10
2
x2
−2
−4
−6
−8
−10
−6 −4 −2 0 2 4 6
x1
Figure 7.16 Phase plane for pendulum under energy control.
104
Solution 7.6
Define V = 21 σ ( x). We then get
V̇ = σ ( x)σ̇ ( x) = ( x1 + x2 )(2x1 + u) = ( x1 + x2 )(−sign)( x1 + x2 ) = −p x1 + x2 p

and therefore σ ( x) = x1 + x2 → 0. The equivalent dynamics is easiest
determine by
d
0= σ ( x) = ẋ1 + ẋ2 = x1 + u + x1
dt
which gives u = −2x1 and hence ẋ1 = − x1 on the sliding plane.
Alternatively one could show that σ ( x) is a sliding mode by noting that
for all x
− x1 − 1
5 6
σ +
f = [1 1] = −1 < 0
x x1
− x1 + 1
5 6
σ −
f = [1 1] =1>0
x x1
However, note the difference compared to showing that σ ( x) is a sliding set,
where the above equations only need to hold for x : σ ( x) = 0.
Solution 7.7
Hamiltonian.
The general form of the Hamiltonian according to Glad/Ljung (18.34) is
H = n 0 ( x2 + u2 ) + λu
Adjoint equation.
λ̇ = − H x = −2n 0 x, λ(1) = free

Optimality conditions.
According to (18.35a) the optimal control signal must minimize the Hamilto-
nian, so the the derivative with respect to u must be zero:
λ
0 = Hu = 2n 0 u + λ [ u=−
2n 0
Hence
λ̇
ẍ = u̇ = −=x
2n 0
The equation ẍ = x has the general solution
x( t) = c 1 e t + c 2 e− t
The boundary conditions x(0) = 1 and x(1) = 0 give
c1 + c2 = 1
c 1 e + c 2 e−1 = 0
This gives c 1 = − e−2 /(1 − e−2 ), c 2 = 1/(1 − e−2 ) and the control signal is
u = ẋ = c 1 e t − c 2 e− t
What about the case n 0 = 0? Then λ is constant and λ(1) = µ ,= 0. Hence
H = λu has no minimum in u, so this case gives no solution candidates.
105
Solution 7.8
Hamiltonian.
φ ( x( t f )) = −0.1x1 ( t f ) − x2 ( t f ) − 5x3 ( t f ) − 3x4 ( t f ) is the criterion to be
minimized. Note that L = 0. Setting α = F /m, we have
H = λ1 x3 + λ2 x4 + λ3 α cos u + λ4 (α sin u − )
Adjoint equation.
λ̇1 = 0
λ̇2 = 0
λ̇3 = −λ1
λ̇4 = −λ2
We know that λ( t f ) = Φ Tx ( x∗ ( t f )) which gives
λ1 ( t f ) = −0.1
λ2 ( t f ) = −1
λ3 ( t f ) = −5
λ4 ( t f ) = −3
Together with the adjoint equation this gives
λ1 ( t) = −0.1
λ2 ( t) = −1
λ3 ( t) = −5 + 0.1( t − t f )
λ4 ( t) = −3 + t − t f
Minimizing H with respect to u gives
H F F
= −λ3 sin u + λ4 cos u = 0
u m m
F F λ4 −3 + t − t f
[ λ3 sin(u) = λ4 cos(u) [ tan(u) = =
m m λ3 −5 + 0.1t − 0.1t f
This gives A = 1, B = −3 − t f , C = 0.1, D = −5 − 0.1t f .
Solution 7.9
We get
u2 u2
H = λ1 x3 + λ2 x4 + λ3 ( cos u1 ) + λ4 ( sin u1 − ) − λ5γ u2
x5 x5
= σ ( t, u1 )u2 + terms independent of u
λ3 λ4
where σ ( t, u1 ) = x5
cos u1 + x5
sin u1 − λ5γ . Since we want to minimize H
with respect to u: 
 umax σ < 0

u2 = ⋆ σ =0


0 σ >0
106
and I
λ4
u2 > 0
tan u1 = λ3
⋆ u2 = 0
Solution 7.10
The problem is normal, can use n 0 = 1. We have
H = e x1 + x22 + u2 + λ1 x2 + λ2 (− x1 − x32 + (1 + x1 )u)

2
2
λ̇1 = − H x1 = −2x1 e x1 − λ2 (−1 + u)
λ̇2 = − H x2 = −2x2 − λ1 + 3x22 λ2
λ(1) = 0
Minimization of H wrt u gives
H λ2
=0 [ 2u + λ2 (1 + x1 ) = 0 [ u=− (1 + x1 )
u 2
2
( uH2 = 2 > 0 hence minimum). This gives
ẋ1 = f1 = x2
λ2
ẋ2 = f2 = − x1 − x32 − (1 + x1 )2
2
2
λ̇1 = f3 = −2x1 e x1 − λ2 (−1 + u)
λ̇2 = f4 = −2x2 − λ1 + 3x22 λ2
λ1 (1) = λ2 (1) = 0
Solution 7.11
Hamiltonian. We can minimize the total time by setting L = 1. No terminal
cost gives φ = 0. The constraint at the final time gives
5 6
x1
Ψ( x) = .
x2
The final time t f is free. The Hamiltonian is given by
H = n 0 + λ1 x2 + λ2 u
Adjoint equation. λ̇( t) = − H T x, λ( t f ) = Ψ T

x ( x) µ
λ̇1 = 0
I 6T 5
1 0 µ1
5 6
, λ( t f ) =
λ̇2 = −λ1 0 1 µ2
Which gives
I
λ1 ( t) = µ1
λ2 ( t) = − µ1 t + B
107
u should be the minimizer for H. Normally we look at Hu , however, this will
not be very useful now as H is linear in u. Instead we note that

 1
 λ2 ( t) < 0
u( t) = arg min H ( x, u, λ, η) = arg min λ2 ( t)u( t) = ? λ2 ( t) = 0
u( t)∈[−1,1] u( t)∈[−1,1] 

−1 λ2 ( t) > 0
Since λ2 is linear, it follows that u( t) = ±1 with at most one switch. The

simplest way to find the switch time is to solve the equations for such input
signals. A common trick is to use dd xx21 to find the trajectories.
d x1 x2 x2
= [ x1 + C1 = 2
d x2 u 2u
For u( t) = 1 we get
x1 + C 1 = x22 /2
This gives the phase plane in the Figure 7.17. For u = −1 we get
x1 + C 2 = − x22 /2
This gives the phase plane in the Figure 7.18. Consider especially the two
5
−1
−2
−3
−4
−5
−5 −4 −3 −2 −1 0 1 2 3 4 5
Figure 7.17 Phase plane for u = 1. The solution is traveling upwards.
curves for u = ±1 that pass through the origin (C 1 = C 2 = 0). We see that
switching has to occur when a curve intersects with another going to the
origin, i.e. when x1 = − 21 sign{ x2 } x22 . To reach the switching curve wee need
u( t) = −1 above the switch curve and u( t) = 1 below. We therefore see that
the control law is given by
1
; <
u( t) = −sign x1 ( t) + sign{ x2 ( t)} x22 ( t) .
2
108
−1
−2
−3
−4
−5
−5 −4 −3 −2 −1 0 1 2 3 4 5
Figure 7.18 Phase plane for u = −1. The solution is traveling downwards.
Phase plane
5
1
x2
−1
−2
−3
−4
−5
−5 −4 −3 −2 −1 0 1 2 3 4
x1
Figure 7.19 Phase plane with switching curve
Solution 7.12
Since we assume the problem is normal (t f is free so this is not obvious) we
have
H = 1 + pup + λ1 x2 + λ2 u.
Minimization wrt pup ≤ 1 gives
λ2 > 1 [ u = −1
pλ2 p < 1 [ u = 0
λ2 < −1 [ u = 1
We also have
λ̇1 = − H x1 = 0 [ λ1 = B
λ̇2 = − H x2 = −λ1 [ λ2 = A − Bt
109
for some constants A, B. If B < 0 we see that λ2 increases (linearly) and hence
u( t) passes through the sequence 1 → 0 → −1, or a subsequence of this. If
B > 0 the (sub-) sequence is passed in the other direction −1 → 0 → −1.
If B = 0 then u is constant: either u = −1, u = 0 or u = 1. The cases
λ2 " 1 and λ2 " −1 are then impossible since the condition H " 0 (since t f
is free) then can not be satisfied.
Solution 7.13
Alternative 1 Use the Bang-bang
 theorem (p. 472). Note that ( A, B) is con-
 1 0 
trollable and Ψ x =  
 has full rank, hence u( t) is bang-bang. From
0 1
“sats 18.6” we know that there are at most n − 1 = 1 switches in u (the
eigenvalues of A are −1, −2 and are hence real).
Alternative 2 Direct calculation shows
H = σ ( t)u + terms independent of u
Minimization wrt u shows that pup = 3 where the sign is given by the sign of
σ ( t). From λ̇ = − AT λ and λ( t f ) = Ψ T
x µ = µ we get
σ ( t) = λT B == µT e− A( t− t f ) B = c 1 e− t + c 2 e−2t
for some constants c 1 , c 2 . Since σ ( t) = e− t (c 1 + c 2 e− t ) can have at most one

sign change and there will be only one switch in u. (It is easy to check that
the case σ ( t) " 0 is impossible).
Solution 7.14
Hamiltonian. Rt
The objective is to minimize t f = 0 f 1d, so L = 1 and the Hamiltonian is
H = n 0 + λT ( Ax + Bu) = λT Bu + λT Ax + n 0
Adjoint equation.
T
λ̇ = − H x = − AT λ [ λ( t) = e− A t λ(0)
The optimal control signal must minimize H, so
u = −sign(λ( t)T B) = −sign(λ(0)T e− At B)
When A = B = 1, this implies that the optimal input is constant, either

u " 1 or u " −1.
110
Solution 7.15
Minimization of
H = (1 + λ1 )u
gives
1 + λ1 ,= 0 : no minimum in u
1 + λ1 = 0 : all u give minima
This does not prove that all u in fact give minima. It only says that all u( t)
are so far possible minima and we need more information.
But in fact since
Z1 Z1
u dt = ẋ1 dt = x(1) − x(0) = 1
0 0
all u that give x1 (1) = 1 are minimizers.
Solution 7.16
(a) Introduce x1 = x, x2 = ẋ
ẋ1 = x2
(7.27)
ẋ2 = − x1 + 2x22 + u − 1
(b) Let ẋ1 = ẋ2 = 0 [ ( x1 , x2 ) = (−1, 0) is the only equilibrium. The

linearization around this point is
0 1 0 1 0
5 6 5 6 5 6
A= = B=
−1 4x2 ( xo1 , xo2 )=(−1, 0) −1 0 1
The characteristic equation for the linearized system is s2 + 1 = 0 [

s = ±i. We can not conclude stability of the nonlinear system from this.
(c) The simplest way is to cancel the constant term and the nonlinearity
with the control signal and introduce some linear feedback.
u = +1 − 2 ẋ2 − a ẋ, a > 0 [ ẍ = − a ẋ − x
As the resulting system is linear and time invariant with poles in the
left half plane for all a > 0 it is GAS.
Solution 7.17
1 2
As the hint suggests, the Lyapunov function V ( x1 , x2 , x3 ) = ( x + x22 + x23 )
2 1
is used:
V (0, 0, 0) = 0, V ( x1 , x2 , x3 ) > 0 for pp xpp ,= 0 and V → +∞ as pp xpp → +∞.
dV
= ẋ1 x1 + ẋ2 x2 + ẋ3 x3 =
dt
− x21 + x1 x2 + x1 x3 tan( x1 ) − x42 − x1 x2 + x3 x22 + ux3 = (7.28)
− x21 − x42 + x3 ( x1 tan( x1 ) + x22 + u)
By choosing e.g. u = − x1 tan( x1 ) − x22 − x3 we will get

dV
dt
= − x21 − x42 − x23 < 0, ∀ x ,= (0, 0, 0). Thus, the closed loop system is GAS
for this choice of u.
111
Solution 7.18
(a) All singular points are given by { ẋ1 = 0, ẋ2 = 0}:
x1 − x2 = 0 and − 3x1 + x31 − x2 = 0 [ x1 = x2 and − 4x1 + x31 = 0
gives x1 = 0, ±2 and x2 = x1
By writing the system with u( t) " 0 and a = 1 as ẋ = f ( x) we get the
linearizations at the equilibria as
f
ẋ ( ( x − x eq )
x p x= xeq
−3 + 3x21 −1
5 6
f
A( x1 , x2 ) = =
x 1 −1
9 −1
5 6
A(2, 2) =
1 −1
√
eig ( A(2, 2)) = 4 ± 24 ( {8.9, −0.9} ( saddle point )
A(−2, −2) gives the same eigenvalues
−3 −1
5 6
A(0, 0) =
1 −1
eig ( A(0, 0)) = {−2, −2} ( stable node )
The origin is only locally asymptotically stable, since there is more than
one equilibrium point. Moreover, solutions starting in the two unstable
equilibrium points will not converge to the origin.
(b)
V̇ = x1 x˙1 + x2 x˙2 = −3x21 + x41 − x1 x2 + x1 x2 < 0
√
as long as p x1 p < 3 and x1 ,= 0. However we see that we can never
prove global stability
√ using this Lyapunov function candidate since
V̇ > 0 if x1 > 3.
Define
1 2
( x + x22 ) ≤ 1.}
Ω = { x1 , x2 : V ( x) =
2 1
Then Ω is an invariant set for the dynamics as (the boundary of) Ω is
a level set for V and it holds that V̇ ≤ 0 on Ω.
As V̇ = 0 for x1 = 0 we must apply LaSalles theorem. The set E where
V̇ = 0 is given by
√ √
E = {( x1 , x2 ) = (0, t), − 2 ≤ t ≤ 2.}
However, when x1 = 0 then x˙1 = − x2 ,= 0 unless x2 also is 0. Therefore

the origin is the only invariant point, and any point starting in Ω
converges to the origin. As the origin is in the interior of Ω we can
conclude local asymptotic stability.
112
(c) If u( x) = − x31 then all nonlinearities are canceled and the system is
purely linear. The eigenvalues are -2,-2 and thus the origin is GAS.
This can showed by using the Lyapunov function as well.
V̇ = −3x21 + x41 − x1 x2 + x1 u + x1 x2 − x22 = −3x21 + x1 ( x31 − u) − x22
= −3x21 − x22 if u( x) = − x31

Then the origin is globally asymptotically stable, since the Lyapunov
function is radially unbounded. The convergence rate is even exponen-
tially since the closed loop system is linear.
Solution 7.19
Both Va and Vb are positive definite with respect to ( x1 , x2 ) and radially
unbounded.
(a)
d
Va = 2 ( x1 ẋ1 + x2 ẋ2 ) = 2( x21 + u) x2
dt
u would need to be u = − x21 − f ( x2 ), where f ( x2 ) is some function that
satisfies I
> 0 x2 > 0
f ( x2 ) =
< 0 x2 < 0,
to get the derivative dV

dt
a
negative (semi-)definite. As u is bounded this
can not be achieved globally.
(b)
x21
3 4 3 4
d x1 ẋ1
Vb = 2 + x2 ẋ2 = 2 + u x2
dt 1 + x21 1 + x21
x21
As 0 ≤ ≤ 1 we can always compensate for this term with u and
1 + x21
by choosing ”the rest of our available control signal” u as for instance
−4sat( x2 ), so that pvp < 5
x21 x21
v=− − 4sat( x2 ) [ u = − − 4sat( x2 )
1 + x21 1 + x21
However, this will leave us with dtd

Vb = −4sat( x2 ) x2 ≤ 0. If x2 = 0 [
ẋ1 = 0, but with the chosen control law ẋ2 = 0 only if x1 = 0, so the
origin will be the only equilibrium.
Solution 7.20
Consider the system
ẋ1 = x2 − x1
ẋ2 = kx21 − x2 + u (7.29)
where u is the input and k is an unknown coefficient.
113
1
We use the Lyapunov function candidate V ( x1 , x2 , k̂) = x21 + x22 + ( k − k̂)2 .,
! "
2
and investigate the time derivative
d
V ( x1 , x2 , k̂) = x1 ẋ1 + x2 ẋ2 − ( k − k̂) k̂˙
dt
since k̇ ∼ 0 because k changes very slowly. Inserting the system equations
and some simplifications gives
d
V ( x1 , x2 , k̂) = x1 x2 − x21 + kx21 x2 − x22 + ux2 − ( k − k̂) k̂˙
dt
= − x21 − x22 + x1 x2 + ux2 + k − k̂˙ + x21 x2 + k̂ k̂˙
1 2
If we chose the update law for the estimate as
k̂˙ = x21 x2 (7.30)
we obtain
d
V ( x1 , x2 , k̂) = − x21 − x22 + x1 x2 + ux2 + k̂x21 x2
dt
= − x21 − x22 + x2 u + x1 + k̂x21
! "
which is now independent of the unknown parameter k. We can now proceed

as usual with the control design.
Choosing u = − x1 − k̂x21
which gives
d
V ( x1 , x2 , k̂) = − x21 − x22
dt
which is negative semi-definite (V ( x1 , x2 , k̂) does not depend on the esti-
mation error). We can not show asymptotically stability, since we can not
guarantee that the estimation error goes to zero. In practice this means that
if the system is at rest, the estimation error will not change, even if the es-
timate is wrong. The estimate is only updated if x1 , x2 ,= 0. This is a general
problem for adaptive systems.
Solution 7.21
Start with the system ẋ1 = x21 + φ ( x1 ) which can be stabilized using φ ( x1 ) =
− x21 − x1 . Notice that φ (0) = 0. Take V1 ( x1 ) = x21 /2. To backstep, define
z2 = ( x2 − φ ( x1 )) = x2 + x21 + x1 ,
to transfer the system into the form
ẋ1 = − x1 + z2
ż2 = u + (1 + 2x1 )(− x1 + z2 )
Taking V = V1 ( x1 ) + z22 /2 as a Lyapunov function gives
V̇ = x1 (− x1 + z2 ) + z2 (u + (1 + 2x1 )(− x1 + z2 )) = − x21 − z22
if u = u = −(1 + 2x1 )(− x1 + z2 ) − x1 − z2 Hence, the origin is globally

114
Solution 7.22
(a) Start with the system ẋ1 = x21 − x31 + φ ( x1 ) which can be stabilized
using φ ( x1 ) = − x21 − x1 . Notice that φ (0) = 0. Take V1 ( x1 ) = x21 /2. To
backstep, define
ζ2 = ( x2 − φ ( x1 )) = x2 + x21 + x1 ,
ẋ1 = − x1 − x31 + ζ2
ζ˙2 = u + (1 + 2x1 )(− x1 − x31 + ζ2 )
Taking V = V1 ( x1 ) + ζ22 /2 as a Lyapunov function gives
V̇ = x1 (− x1 − x31 + ζ2 ) + ζ2 (u + (1 + 2x1 )(− x1 − x31 + ζ2 )) = − x21 − x41 − ζ22
if u = −(1 + 2x1 )(− x1 + ζ2 ) − x1 − ζ2 = 2x41 − x31 − 2x1 − 2x2 − 2x1 x2 .

Hence, the origin is globally asymptotically stable. Notice that we did
not have to cancel out the term − x31 since it contributes to stability.
(b) The phase plane plot of the system is shown in Figure 7.20
Figure 7.20 Phase plane for system in exercise 7.22.
Solution 7.23
(a) Defining
f1 ( x1 ) = x1
1 ( x1 ) = 1
f2 ( x1 , x2 ) = sin( x1 − x2 )
2 ( x1 , x2 ) = 1
the system can be written on the strict feedback form
ẋ1 = f1 ( x1 ) + 1 ( x1 ) x2
ẋ2 = f2 ( x1 , x2 ) + 2 ( x1 , x2 )u
(see lecture 8).
115
(b) Start with the system ẋ1 = x1 + φ ( x1 ) which can be stabilized using
φ ( x1 ) = −2x1 . Notice that φ (0) = 0. Take V1 ( x1 ) = x21 /2. To backstep,
define
ζ2 = ( x2 − φ ( x1 )) = x2 + 2x1 ,
ẋ1 = − x1 + ζ2
ζ˙2 = −2x1 + 2ζ2 + sin(3x1 − ζ2 ) + u
V̇ = x1 ẋ1 + ζ2ζ˙2 = x1 (− x1 + ζ2 ) + ζ2 (u − 2x1 + 2ζ2 + sin(3x1 − ζ2 ))

= − x21 − ζ22
if u = − sin(3x1 − ζ2 ) + x1 − 3ζ2 = − sin( x1 − x2 ) − 5x1 − 3x2 . Hence,

the origin is globally asymptotically stable.
Solution 7.24
(a) Defining
f1 ( x1 ) = −sat( x1 )
1 ( x1 ) = x21
f2 ( x1 , x2 ) = x21
2 ( x1 , x2 ) = 1
the system can be written on the strict feedback form
ẋ1 = f1 ( x1 ) + 1 ( x1 ) x2
ẋ2 = f2 ( x1 , x2 ) + 2 ( x1 , x2 )u
(see lecture 8).

(b) Start with the system ẋ1 = −sat( x1 ) + x21φ ( x1 ) which can be stabilized
using φ ( x1 ) = − x1 . Notice that φ (0) = 0. Take V1 ( x1 ) = x21 /2. To
backstep, define
ζ2 = ( x2 − φ ( x1 )) = x2 + x1 ,
ẋ1 = −sat( x1 ) − x31 + x21ζ2

ζ˙2 = −sat( x1 ) − x31 + x21ζ2 + x21 + u
Notice that we did not have to cancel out the term −sat( x1 ) since it
contributes to stability.
= − x1 sat( x1 ) − x41 + x31ζ2 + ζ2 −sat( x1 ) − x31 + x21ζ2 + x21 + u

! "
V̇
= − x1 sat( x1 ) − x41 + ζ2 −sat( x1 ) + x21ζ2 + x21 + u
! "
= − x1 sat( x1 ) − x41 − ζ22
if u = sat( x1 ) − x21ζ2 − x21 − ζ2 = sat( x1 ) − x31 − x21 − x1 − x2 − x21 x2 .

Hence, the origin is globally asymptotically stable.
116
Solution 7.25
Start with the system ẋ1 = x1 +φ 1 ( x1 ) which can be stabilized using φ 1 ( x1 ) =
−2x1 . Notice that φ 1 (0) = 0. Take V1 ( x1 ) = x21 /2. To backstep, define
ζ2 = ( x2 − φ 1 ( x1 )) = x2 + 2x1 ,
ẋ1 = − x1 + ζ2
ζ˙2 = −2x1 + 2ζ2 + sin(3x1 − ζ2 ) + x3
Think of x3 as a control input and consider the system
ẋ1 = − x1 + ζ2
ζ˙2 = −2x1 + 2ζ2 + sin(3x1 − ζ2 ) + φ 2
and use the Lyapunov function candidate V2 = V1 + ζ22 /2:
V̇2 = x1 ẋ1 + ζ2ζ˙2 = x1 (− x1 + ζ2 ) + ζ2 (−2x1 + 2ζ2 + sin(− x1 + 2ζ2 ) + φ 2 )

= − x21 − ζ22 + ζ2 (− x1 + 3ζ2 + sin(3x1 − ζ2 ) + φ 2 )
| {z }
which is stabilized by
φ2 = − sin(3x1 − ζ2 ) + x1 − 3ζ2
W
V̇2 = − x21 − ζ22
To backstep a second time, define
ζ3 = x3 − φ 2 = x3 + sin(3x1 − ζ2 ) − x1 + 3ζ2
=[
ζ˙3 = ẋ3 + cos(3x1 − ζ2 ) · (3 ẋ1 − ζ˙2 ) − ( ẋ1 + 3ζ˙2 )
= u + cos(3x1 − ζ2 )(−2x1 + 4ζ2 − ζ3 ) − 2x1 − 4ζ2 + 3ζ3
ẋ1 = − x1 + ζ2
ζ˙2 = − x1 − ζ2 + ζ3
ζ˙3 = u + cos(3x1 − ζ2 )(−2x1 + 4ζ2 − ζ3 ) − 2x1 − 4ζ2 + 3ζ3
= u + β ( x, z)
Now the control signal appears in the equation, and we can design a control
law. Consider the Lyapunov function candidate V = V2 + ζ32 /2:
V̇ = x1 ẋ1 + ζ2ζ˙2 + ζ3ζ˙3

= x1 (− x1 + ζ2 ) + ζ2 (− x1 − ζ2 + ζ3 ) + ζ3 (u + β ( x1 , ζ2 , ζ3 ))
= − x21 − ζ22 − ζ32 + ζ3 (ζ2 + ζ3 + u + β ( x1 , ζ2 , ζ3 ))
| {z }
117
Choosing
u = −ζ2 − ζ3 − β = cos(3x1 − ζ2 )(−2x1 + 4ζ2 − ζ3 ) + 2x1 + 3ζ2 − 4ζ3
gives
V̇ = − x21 − ζ22 − ζ32
which is negative definite, and the system is therefore global asymptotically

stable
Solution 7.26
(a) Define
N
X −1
Vk ( x k ) = N ( x N ) + j ( x j , µ∗j ( x j ))
j= k
where µ∗ = { µ0 , µ1 , . . . , µ N −1 } is an optimal control policy. Following

∗ ∗ ∗
the lecture notes, we have
V4 ( x) = 4 ( x) = x2
V3 ( x) = min [ 3 ( x, u) + V4 (2x + u)]
u
= min x2 + u2 + (2x + u)2 = 3x2 (the minimum is attained for u = − x)

# $
u
V2 ( x) = min [ 2 ( x, u) + V3 (2x + u)]
u
= min x2 + 3u2 + 3(2x + u)2 = 7x2 (the minimum is attained for u = − x)

# $
u
V1 ( x) = min [ 1 ( x, u) + V2 (2x + u)]
u
= min x2 + 7u2 + 7(2x + u)2 = 15x2 (the minimum is attained for u = − x)

# $
u
V0 ( x) = min [ 0 ( x, u) + V1 (2x + u)]
u
= min x2 + 15u2 + 15(2x + u)2 = 31x2 (the min is attained for u = − x)

# $
u
The minimal values are attained by the control law u k = − x k for

k = 0, 1, 2, 3.
(b) The optimal control law generates the sequence
u3 = u2 = u1 = u0 = 2.
and the optimal value V0 ( x0 ) = 31x20 = 124.

(c) The control constraint pu0 p ≤ 1 changes only the last step in the dynamic
programming sequence:
V0 ( x) = min x2 + 15u2 + 15(2x + u)2

# $
pup≤1
= min 31x2 + 30( x + u)2

# $
pup≤1
31x2 if p xp ≤ 1
I
=
31x2 + 30( x − sgn( x))2 otherwise
118
Solution 7.27
The Hamiltonian-Jacobi-Bellman equation for this problem is
Vt ( t, x) = −min[ L +∇ x V f ] \ q̇x2 = −min[ x2 u2 +2qx2 u] = −min[u2 +2qu] x2
u u u
Minimization yields u = −q. Insertion gives
q̇x2 = −(q2 − 2q2 ) x2 = q2 x2 \ q̇ = q2
Solving the differential equation yields q = C1− t , where C is a constant. It

remains to determine C. For this, we use that the final "cost-to-go" is given
by
V (1, x) = Φ( x) = x2 \ q(1) x2 = x2 \ q(1) = 1.

1
This yields C = 2. Since u = −q, the optimal control is u( t, x) = t−2 .
119
8. Bibliography
Åström, K. J. (1968): Reglerteknik – Olinjära System. TLTH/VBV.
Boyd, S. P. (1997): “Homework assignments in ee375 – advanced analysis of
feedback.” Available from http://www-leland.stanford.edu/class/ee375/.
Khalil, H. K. (1996): Nonlinear Systems, 2nd edition. Prentice Hall, Upper
Saddle River, N.J.
Slotine, J.-J. E. and W. LI (1991): Applied Nonlinear Control. Prentice Hall,
Englewood Cliffs, N.J.
120

Compendium Solutions NLcontrol

Uploaded by

Copyright:

Available Formats

Compendium Solutions NLcontrol

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Compendium Solutions NLcontrol

Uploaded by

Copyright:

Available Formats

Exercises in

Nonlinear Control Systems

The authors, Jan 1999

mlθ¨ = −m  sin θ − klθ˙,

Figure 1.1 The pendulum in Exercise 1.1

Exercise 1.2[Khalil, 1996]

where q 1 and q 2 are angular positions, I and J are moments of inertia, k is a

Figure 1.2 The flexible manipulator in Exercise 1.2

Exercise 1.3[Khalil, 1996]

where δ is the rotor deflection angle in radians, E q is voltage, P is mechanical

Figure 1.3 The feedback system in Exercise 1.4

Figure 1.3 shows a feedback connection of a linear time-invariant system and

Exercise 1.5[Khalil, 1996]

Figure 1.4 The phase-locked loop in Exercise 1.5

A phase-locked loop can be represented by the block diagram of Figure 1.4.

is a state equation for the closed-loop system.

Exercise 1.6 (H)

Figure 1.5 Control system with friction in Example 1.6.

Figure 1.6 Anti-windup compensation in Example 1.7.

Motor Valve Process

Figure 1.7 Block diagram of system in Exercise 1.8.

that the valve characteristic is given by f ( x) = x2 (this is unrealistic for

Is the system nonlinear locally controllable at (0, 0, 0, 0)?

ẋ1 = cos( x1 ) − 1 + x22 + u

Figure 1.10 The flow system in Exercise 1.15

Figure 1.11 The inverted pendulum on an electric hand-saw in Exercise 1.16

(c) ẋ1 = (1 − x1 ) x1 − 2x1 x2 /(1 + x1 )

(f) ẋ1 = ( x1 − x2 )( x21 + x22 − 1)

Exercise 2.2[Åström, 1968]

and determine the type of equilibrium.

Exercise 2.3[Khalil, 1996]

(c) ẋ1 = 2x1 − x1 x2

Which phase portrait belongs to what system?

Phase plane Phase plane Phase plane

Exercise 2.5[Khalil, 1996] (H)

Phase plane Phase plane

ẋ1 = (u − x1 )(1 + x22 )

is controlled by the output feedback

Exercise 2.7[Åström, 1968]

The equilibrium points of the system where derived in Exercise 1.3(c).Determine

Exercise 2.8 (H)

ẋ1 = x2 + x1 (1 − x21 − x22 )

(a) Verify that the trajectory ( x1 , x2 ) = (sin t, cos t) is a solution to the

around the trajectory

Linearize the equation around this nominal solution.

Exercise 2.8 Introduce polar coordinates to determine stability of the limit

(a) Show that Lyapunov’s linearization method fails to determine stability

to show that the system is globally asymptotically stable for a < 0.

where d > 0, k > 0. Show that

where Q = QT is any positive definite matrix. The system is asymptot-

Verify by completing squares that V ( x) = xT P x is a positive

p 11 > 0, p 11 p 22 − p212 > 0

(c) Solve the Lyapunov equation in Matlab.

Exercise 3.5[Slotine and LI, 1991]

ẋ( t) = Ax( t), t ≥ 0,

is asymptotically stable if all eigenvalues of A have negative real parts. It

ẋ( t) = A( t) x( t), t ≥ 0, (3.1)

is asymptotically stable if the eigenvalues of A( t) have negative real parts

(b) The system (3.1) is however stable if the eigenvalues of A( t) + AT ( t)

mlθ¨ = −m sin θ − klθ˙,

Let the nonlinearity be on the form ( e) = e3 and investigate stability of the