Outline

Lyapunov Stability Theory: Linear  Lyapunov’s (first, indirect) linearization

Systems method.
 Linear time-invariant case.
M. S. Fadali
Professor of EE  Domain of attraction.

1 2

Lyapunov’s Linearization Method Example

 Linearize nonlinear system in vicinity  Determine the stability of the equilibrium
of equilibrium : of the mechanical system at the origin
.
 Find the eigenvalues of the linearized system.
The equilibrium of the nonlinear system is:  Equilibrium with
◦ Exponentially stable if all the eigenvalues are in the
open LHP.
◦ Unstable if one or more of its eigenvalues is in the
open RHP.
◦ Inconclusive for LHP eigenvalues and one or
more eigenvalues on the imaginary axis.
3 4
Nonlinear State Equations Linearization and Stability
 Physical state variables  Equilibrium state
 Linearized model with
 State Equations
 Characteristic polynomial and stability

 Stable ,

5 6

Linear Time-invariant Case Proof: Sufficiency

The LTI system  Use a quadratic Lyapunov function

is asymptotically stable if and only if for

any positive definite matrix there exists
a positive definite symmetric solution to
the Lyapunov equation

globally exp. stable.

7 8
 Proof: Necessity Symmetric Positive Definite
 Let Hurwitz

for some nonzero

iff is not an observable pair.
for observable.
→
Note: can be positive semidefinite.
9 10

Uniqueness Remarks
 Recall that the original Lyapunov theorem
only gives a sufficient condition.
Subtract  If we start with (i.e. with ) and
solve for , the condition the test may or
may not work.
 If we start with (i.e. with the derivative
and we find a the condition is necessary
constant if and only if and sufficient.

11 12
Example Solution
Determine the stability of the system with
state matrix

• Multiply
using the Lyapunov equation with .
Note: The system is clearly stable by
inspection since is in companion form.

• Equate to obtain three equations in three unknowns.

13 14

Equivalent Linear System

Choose

 not positive definite.

 No conclusion: sufficient condition only.
 Choose and solve for .

15 16
MAPLE Equivalent Linear System
Compute:
with(LinearAlgebra):
Transpose(A).P+P.A

Solve the equivalent linear system: M.p=-q

p is a vector whose entries are the entries
of the P matrix, similarly define q
LinearSolve(M,B)

17 18

MATLAB MATLAB Example

>> A=[0,1;-6,-5];
>> Q=eye(2)
 Solve a different equation. >> P=lyap(A,eye(2))
P=
 Identical to our equation with 0.5333 -0.5000
replaced by . -0.5000 0.7000
>> eig(P)
ans =
 Eigenvalues are the same! 0.1098
1.1236

19 20
 To Get Earlier Answer Domain (Ball, Region) of Attraction
 1.1167 0.08333  Region in which the trajectories of the
P 
0.08333 0.1167  system converge to an asymptotically
stable equilibrium point.
>> P=lyap(A',eye(2))  Difficult to estimate, in general.
P=  Can be estimated using the linearized
1.1167 0.0833 system in the vicinity of the asymptotically
0.0833 0.1167 stable equilibrium.

21 22

Example Calculate

Equilibrium

Lyapunov function candidate for

For

23 24
Simulation Results Theorem 3.9
 The ball of attraction can be estimated  Equilibrium of
to be
 Although for I. compact set containing ,
we have this invariant w.r.t. the solutions of
region includes divergent trajectories II.
because is not an invariant set. For
example, the trajectory starting at Then the region of attraction of
crosses then
diverges.
25 26

Proof Example
 Under the assumptions

 is the largest invariant set in

 By La Salle’s Theorem, every solution
starting in approaches as , i.e.
approaches as For
 is an estimate of the domain of
attraction.

27 28
 Invariant Set Estimate Using Linearized system

Minimum value at edge

Solve

29 30

Example Contours
2

Equilibrium 1.5

Solve 1

0.5

-0.5

-1

-1.5
-3 -2 -1 0 1 2 3

for
31 32