Introduction

Linearization
Solution to ordinary di↵erential equations
Lyapunov Stability Theory
Observer design

Remarks: Some coordinates may be more convenient than others

Nonlinear system with X = R2 and x = (x1 , x2 )

ẋ1 = x2 x12
⇣ ⌘
ẋ2 = x1 + 2x1 x2 x12 +u

Consider the di↵eomorphism : X 7! X given by

  
z1 x1 1 z1
z= = (x) = ,x = (z) =
z2 x2 x12 z12 + z2

Same system in new coordinates z :

M. Nadri Control of nonlinear dynamic systems 1 36 / 105

 Introduction
Linearization
Solution to ordinary di↵erential equations
Lyapunov Stability Theory
Observer design

Remarks: Some coordinates may be more convenient than others

Nonlinear system with X = R2 and x = (x1 , x2 )

ẋ1 = x2 x12
⇣ ⌘
ẋ2 = x1 + 2x1 x2 x12 +u

Consider the di↵eomorphism : X 7! X given by

  
z1 x1 1 z1
z= = (x) = ,x = (z) =
z2 x2 x12 z12 + z2

Same system in new coordinates z :

ż1 = z2
ż2 = z1 + u
the system becomes linear when written in the new coordinates!

M. Nadri Control of nonlinear dynamic systems 1 36 / 105

 Introduction
Eequilibrium Points
Linearization
Phase Plane Analysis
Lyapunov Stability Theory
Nonlinear Phenomena
Observer design

Equilibrium Points

A point x = x ⇤ in the state space is said to be an equilibrium point of

ẋ = f (t, x) if

x (t0 ) = x ⇤ ) x(t) ⌘ x ⇤ , 8t t0
For the autonomous system x = f (x), the equilibrium points are the real
solutions of the equation
f (x) = 0

An equilibrium point could be isolated; that is, there are no continuum of

equilibrium points.

M. Nadri Control of nonlinear dynamic systems 1 37 / 105

 Introduction
Eequilibrium Points
Linearization
Phase Plane Analysis
Lyapunov Stability Theory
Nonlinear Phenomena
Observer design

Linear system ẋ = Ax can have an isolated equilibriumpoint at x = 0 (if A

is nonsingular) or a continuum of equilibrium points in the null space of A
(if A is singular)

Linear system cannot have multiple isolated equilibrium points, if xa and

xb are two equilibrium points, then by linearity any point on the form
↵xa + (1 ↵)xb connecting xa and xb will be an equilibrium point

Nonlinear system can have multiple isolated equilibrium points. For

example, the state equation:

ẋ1 = x2 , ẋ2 = a sin x1 bx2

M. Nadri Control of nonlinear dynamic systems 1 38 / 105

 Introduction
Eequilibrium Points
Linearization
Phase Plane Analysis
Lyapunov Stability Theory
Nonlinear Phenomena
Observer design

Linear system ẋ = Ax can have an isolated equilibriumpoint at x = 0 (if A

is nonsingular) or a continuum of equilibrium points in the null space of A
(if A is singular)

Linear system cannot have multiple isolated equilibrium points, if xa and

xb are two equilibrium points, then by linearity any point on the form
↵xa + (1 ↵)xb connecting xa and xb will be an equilibrium point

Nonlinear system can have multiple isolated equilibrium points. For

example, the state equation:

ẋ1 = x2 , ẋ2 = a sin x1 bx2

has equilibrium points at (x1 = n⇡, x2 = 0) for n = 0, ±1, ±2, · · ·

M. Nadri Control of nonlinear dynamic systems 1 38 / 105

 Introduction
Eequilibrium Points
Linearization
Phase Plane Analysis
Lyapunov Stability Theory
Nonlinear Phenomena
Observer design

Linearization

A common engineering practice in analyzing a nonlinear system is to linearize it

about some nominal operating point and analyze the resulting linear model.

What are the limitations of linearization?

Since linearization is an approximation in the neighborhood of an
operating point, it can only predict the ”local” behavior of the nonlinear
system in the vicinity of that point. It cannot predict the ”nonlocal” or
”global” behavior
There are ”essentially nonlinear phenomena” that can take place only in
the presence of nonlinearity.

M. Nadri Control of nonlinear dynamic systems 1 39 / 105

 Introduction
Eequilibrium Points
Linearization
Phase Plane Analysis
Lyapunov Stability Theory
Nonlinear Phenomena
Observer design

Linearization about a trajectory

Linearization about a trajectory Given: a specific trajectory (x̄(t), ū(t), ȳ (t))

˙
x̄(t) = f (t, x̄(t), ū(t))
ȳ (t) = h(t, x̄(t), ū(t))

Define: error variables

x(t) = x(t) x̄(t), u(t) = u(t) ū(t), y (t) = y (t) ȳ (t)

Compute: linear approximation from Jacobian matrices

 
@f @f
ẋ(t) ⇡ (t, x̄(t), ū(t) x(t) + (t, x̄(t), ū(t) u(t).
@x @u
 
@h @h
y (t) ⇡ (t, x̄(t), ū(t) x(t) + (t, x̄(t), ū(t)] u(t).
@x @u

M. Nadri Control of nonlinear dynamic systems 1 40 / 105

 Introduction
Eequilibrium Points
Linearization
Phase Plane Analysis
Lyapunov Stability Theory
Nonlinear Phenomena
Observer design

Linearization about an equilibrium trajectory

An equilibrium is a constant trajectory (x̄, ū, ȳ ) and it satisfies 0 = f (t, x̄, ū)
and ȳ = h(t, x̄, ū) for all t.

On the Simple pendulum example

M. Nadri Control of nonlinear dynamic systems 1 41 / 105

 Introduction
Eequilibrium Points
Linearization
Phase Plane Analysis
Lyapunov Stability Theory
Nonlinear Phenomena
Observer design

Linearization about an equilibrium trajectory

An equilibrium is a constant trajectory (x̄, ū, ȳ ) and it satisfies 0 = f (t, x̄, ū)
and ȳ = h(t, x̄, ū) for all t.

On the Simple pendulum example

Equilibrium trajectory with ū = mg ` sin (x̄1 ) and x̄2 = 0 :
 
0 1 0
ẋ ⇡ g b x+ 1 u
`
cos (x̄1 ) m`2 m`2

M. Nadri Control of nonlinear dynamic systems 1 41 / 105

 Introduction
Eequilibrium Points
Linearization
Phase Plane Analysis
Lyapunov Stability Theory
Nonlinear Phenomena
Observer design

Linearization about an equilibrium trajectory

An equilibrium is a constant trajectory (x̄, ū, ȳ ) and it satisfies 0 = f (t, x̄, ū)
and ȳ = h(t, x̄, ū) for all t.

On the Simple pendulum example

Equilibrium trajectory with ū = mg ` sin (x̄1 ) and x̄2 = 0 :
 
0 1 0
ẋ ⇡ g b x+ 1 u
`
cos (x̄1 ) m`2 m`2

The LTI system is:

- controllable
- stable when the equilibrium angle x̄1 is in the bottom half of the circle, that is
when |x̄1 | < ⇡2
- unstable when the x̄1 is in the top half of the circle

M. Nadri Control of nonlinear dynamic systems 1 41 / 105

 Introduction
Eequilibrium Points
Linearization
Phase Plane Analysis
Lyapunov Stability Theory
Nonlinear Phenomena
Observer design

Example 1

Consider the following system

⇢
ẋ1 = f1 (x1 , x2 , u) = C1 u x1 k (u x1 )3 x2
ẋ2 = f2 (x1 , x2 , u) = L1 x1

where C > 0, L > 0.

Give the linearized system for the equilibrum point x̄ = [0, 0], ū = 0.

M. Nadri Control of nonlinear dynamic systems 1 42 / 105

 Introduction
Eequilibrium Points
Linearization
Phase Plane Analysis
Lyapunov Stability Theory
Nonlinear Phenomena
Observer design

Example 1

Consider the following system

⇢
ẋ1 = f1 (x1 , x2 , u) = C1 u x1 k (u x1 )3 x2
ẋ2 = f2 (x1 , x2 , u) = L1 x1

where C > 0, L > 0.

Give the linearized system for the equilibrum point x̄ = [0, 0], ū = 0.
" # 
@f1 @f1
@x1 @x2
1
+ 3k (u x1 )2 1
J (x1 , x2 , u) = @f2 @f2 = C C
1
C
@x @x L
0
1 2

which gives  1 1
J (x̄1 , x̄2 , ū) = A = C C
1
L
0
2 1 1
The characteristic equation is + C
+ LC
= 0.

M. Nadri Control of nonlinear dynamic systems 1 42 / 105

 Introduction
Eequilibrium Points
Linearization
Phase Plane Analysis
Lyapunov Stability Theory
Nonlinear Phenomena
Observer design

Example 2

Let consider system :

⇢
ẋ1 = x2 + ✏x1 x12 + x22
ẋ2 = x1 + ✏x2 x12 + x22

where ✏ 6= 0 and the origin is an equilibrium point.

Give the linearized system:

M. Nadri Control of nonlinear dynamic systems 1 43 / 105

 Introduction
Eequilibrium Points
Linearization
Phase Plane Analysis
Lyapunov Stability Theory
Nonlinear Phenomena
Observer design

Example 2

Let consider system :

⇢
ẋ1 = x2 + ✏x1 x12 + x22
ẋ2 = x1 + ✏x2 x12 + x22

where ✏ 6= 0 and the origin is an equilibrium point.

Give the linearized system:

3✏x12 + ✏x22 1 + 2✏x1 x2
Jx =
1 + 2✏x1 x2 ✏x12 + 3✏x22

evaluated on x1 = 0, x2 = 0 :

0 1
A=
1 0

which has two complex eignevalues 1,2 = ±i.

M. Nadri Control of nonlinear dynamic systems 1 43 / 105

 Introduction
Eequilibrium Points
Linearization
Phase Plane Analysis
Lyapunov Stability Theory
Nonlinear Phenomena
Observer design

Phase portrait

M. Nadri Control of nonlinear dynamic systems 1 44 / 105

 Introduction
Eequilibrium Points
Linearization
Phase Plane Analysis
Lyapunov Stability Theory
Nonlinear Phenomena
Observer design

Phase portrait

ẋ1 = f1 (x1 , x2 ) = f1 (x)

ẋ2 = f2 (x1 , x2 ) = f2 (x)
Let x(t) = (x1 (t), x2 (t)) be a solution that starts at initial state x0 = (x10 , x20 ).
The locus in the x1 x2 plane of the solution x(t) for all t 0 is a curve
that passes through the point x0 .
This curve is called a trajectory or orbit
The x1 x2 plane is called the phase plane
The family of all trajectories is called the phase portrait
The vector field f (x) = (f1 (x), f2 (x)) is tangent to the trajectory at point
x because
dx2 f2 (x)
=
dx1 f1 (x)

M. Nadri Control of nonlinear dynamic systems 1 45 / 105

 Introduction
Eequilibrium Points
Linearization
Phase Plane Analysis
Lyapunov Stability Theory
Nonlinear Phenomena
Observer design

Numerical Construction of the Phase Portrait

Select a bounding box in the state plane

Select an initial point x0 and calculate the trajectory through it by solving

ẋ = f (x), x(0) = x0

Repeat the process interactively

M. Nadri Control of nonlinear dynamic systems 1 46 / 105

 Introduction
Eequilibrium Points
Linearization
Phase Plane Analysis
Lyapunov Stability Theory
Nonlinear Phenomena
Observer design

Numerical Construction of the Phase Portrait

Select a bounding box in the state plane

Select an initial point x0 and calculate the trajectory through it by solving

ẋ = f (x), x(0) = x0

Repeat the process interactively

[0.5; 1]; [-0.5;-1] [-1; 1]; [1;-1] ẋ1 = 2x1 + x2 ; ẋ2 = 2x1 + 3x2

M. Nadri Control of nonlinear dynamic systems 1 46 / 105

 Introduction
Eequilibrium Points
Linearization
Phase Plane Analysis
Lyapunov Stability Theory
Nonlinear Phenomena
Observer design

Qualitative Behavior of Linear Systems

ẋ = Ax, A is a 2 ⇥ 2 real matrix
1
x(t) = M exp (Dt) M x0

When A has distinct eigenvalues,

 
1 0 ↵
D= or
0 2 ↵
x(t) = Mz(t)
ż = Dz(t)

Case 1: Both eigenvalues are real: M = [v1 , v2 ]

v1 , v2 are the real eigenvectors associated with 1 , 2

ż1 = 1 z1 , ż2 = 2 z2
1t 2t
z1 (t) = z10 e , z2 (t) = z20 e
/ 1 2/ 1
z2 = cz1 2 , c = z20 / (z10 ) .

The shape of the phase portrait depends on the signs of 1 and 2.

M. Nadri Control of nonlinear dynamic systems 1 47 / 105
 Introduction
Eequilibrium Points
Linearization
Phase Plane Analysis
Lyapunov Stability Theory
Nonlinear Phenomena
Observer design

2 < 1 <0

1t 2t
e and e tend to zero as t ! 1
2t 1t
e tends to zero faster than e
Call 2 the fast eigenvalue ( v2 the fast eigenvector) and 1 the slow
eigenvalue (v1 the slow eigenvector)
/ 1
The trajectory tends to the origin along the curve z2 = cz1 2 with
2/ 1 > 1
dz2 2 [( / ) 1]
= c z1 2 1
dz1 1

M. Nadri Control of nonlinear dynamic systems 1 48 / 105

 Introduction
Eequilibrium Points
Linearization
Phase Plane Analysis
Lyapunov Stability Theory
Nonlinear Phenomena
Observer design

M. Nadri Control of nonlinear dynamic systems 1 49 / 105

 Introduction
Eequilibrium Points
Linearization
Phase Plane Analysis
Lyapunov Stability Theory
Nonlinear Phenomena
Observer design

M. Nadri Control of nonlinear dynamic systems 1 50 / 105

 Introduction
Eequilibrium Points
Linearization
Phase Plane Analysis
Lyapunov Stability Theory
Nonlinear Phenomena
Observer design

<0< 1
2
1t 2t
e ! 1 while e tend to zero as t ! 1
Call 2 the stable eigenvalue ( v2 the stable eigenvector) and 1 the
unstable eigenvalue (v1 the unstable eigenvector)
/
z2 = cz1 2 1 with 2 / 1 < 1

M. Nadri Control of nonlinear dynamic systems 1 51 / 105

 Introduction
Eequilibrium Points
Linearization
Phase Plane Analysis
Lyapunov Stability Theory
Nonlinear Phenomena
Observer design

Case 2. Complex eigenvalues: 1,2 = ↵ ± j

We can show that there exists a coordinate such that :
 
1 ↵ cos( t) sin( t)
J = M AM = , exp(Jt) = exp(↵t)
↵ sin( t) cos( t)
ż1 = ↵z1 z2 , ż2 = z1 + ↵z2
q ✓ ◆
1 z2
r = z12 + z22 , ✓ = tan
z1
r (t) = r0 e ↵t and ✓(t) = ✓0 + t
↵ < 0 ) r (t) ! 0 as t ! 1
↵ > 0 ) r (t) ! 1 as t ! 1
↵ = 0 ) 8t, r (t) ⌘ r0 .

M. Nadri Control of nonlinear dynamic systems 1 52 / 105

 Introduction
Eequilibrium Points
Linearization
Phase Plane Analysis
Lyapunov Stability Theory
Nonlinear Phenomena
Observer design

Case 2. Complex eigenvalues: 1,2 = ↵ ± j

We can show that there exists a coordinate such that :
 
1 ↵ cos( t) sin( t)
J = M AM = , exp(Jt) = exp(↵t)
↵ sin( t) cos( t)
ż1 = ↵z1 z2 , ż2 = z1 + ↵z2
q ✓ ◆
1 z2
r = z12 + z22 , ✓ = tan
z1
r (t) = r0 e ↵t and ✓(t) = ✓0 + t
↵ < 0 ) r (t) ! 0 as t ! 1
↵ > 0 ) r (t) ! 1 as t ! 1
↵ = 0 ) 8t, r (t) ⌘ r0 .

M. Nadri Control of nonlinear dynamic systems 1 52 / 105

 Introduction
Eequilibrium Points
Linearization
Phase Plane Analysis
Lyapunov Stability Theory
Nonlinear Phenomena
Observer design

Qualitative Behavior Near Equilibrium Points

The qualitative behavior of a nonlinear system near an equilibrium point can

take one of the patterns we have seen with linear systems. Correspondingly the
equilibrium points are classified as stable node, unstable node, saddle, stable
focus, unstable focus, or center.

M. Nadri Control of nonlinear dynamic systems 1 53 / 105

 Introduction
Eequilibrium Points
Linearization
Phase Plane Analysis
Lyapunov Stability Theory
Nonlinear Phenomena
Observer design

Qualitative Behavior Near Equilibrium Points

The qualitative behavior of a nonlinear system near an equilibrium point can

take one of the patterns we have seen with linear systems. Correspondingly the
equilibrium points are classified as stable node, unstable node, saddle, stable
focus, unstable focus, or center.
Can we determine the type of the equilibrium point of a nonlinear system by
linearization?

M. Nadri Control of nonlinear dynamic systems 1 53 / 105

 Introduction
Eequilibrium Points
Linearization
Phase Plane Analysis
Lyapunov Stability Theory
Nonlinear Phenomena
Observer design

Qualitative Behavior Near Equilibrium Points

The qualitative behavior of a nonlinear system near an equilibrium point can

take one of the patterns we have seen with linear systems. Correspondingly the
equilibrium points are classified as stable node, unstable node, saddle, stable
focus, unstable focus, or center.
Can we determine the type of the equilibrium point of a nonlinear system by
linearization?
Type of equilibrium point
Eigenvalues of A
of the nonlinear system
2 < 1 < 0 Stable Node
2 > 1 > 0 Unstable Node
2 < 0 < 1 Saddle
↵ ± j ,↵ < 0 Stable Focus
↵ ± j ,↵ > 0 Unstable Focus
±j Linearization Fails

M. Nadri Control of nonlinear dynamic systems 1 53 / 105