You are on page 1of 7

Stability Definitions

Stability, Detectability & Stability: in the sense of Lyapunov (i.s. L).


Asymptotic (Internal) Stability: Zero-input
Stabilizability response.
Input-Output Stability or Bounded-Input-
M. Sami Fadali Bounded-Output (BIBO) Stability: Zero-
Professor of Electrical Engineering state response.
University of Nevada

1 2

Equilibrium State Stability of Equilibrium


How sensitive is the system to small
perturbations in its equilibrium?
Solve Unstable
= equilibrium state
For nonlinear systems, multiple states Stable

Stability of equilibrium state: depends on


behavior after a perturbation from the
equilibrium. Asymptotically
stable
3 4
Stability Stable System
Definition: For any there exists a
constant such that implies

Can stay arbitrarily close to equilibrium by 0

starting sufficiently close to it.


Unstable: not stable (cannot stay arbitrarily
close to the equilibrium.
: = open ball of radius
: = open ball of radius
5 6

Exponential Stability
Asymptotically Stable
Definition: There exist positive constants
Definition: Stable equilibrium and it is such that implies
possible to choose such that
implies Global exponential stability: property holds

Converges to equilibrium by starting for any initial state .
sufficiently close to it. Length of state vector decays faster than an
Globally asymptotically stable: converges exponential function.
to equilibrium from any initial state. For linear systems, decay is always
exponential.
7 8
Exponential Stability Linear Time-invariant Systems
1

0.9
For a nonsingular state matrix
0.8

0.7
if and only if
0.6 Only one equilibrium point at the origin.
0.5
For a singular state matrix
0.4

0.3
,
0.2 Rank deficit=number of linearly independent
0.1
Infinitely many equilibrium points on
0
0 5 10 15

9 10

Asymptotic (Internal) Stability for


Stability for LTI Systesm LTI Systems
Theorem: LTI system is asymptotically stable if
the zero-input response converges to zero for any
initial state.

Response is bounded for distinct eigenvalues on LTI system is asymptotically stable if and only if
the imaginary axis (stable). all eigenvalues are in the open LHP.
Response is unbounded for repeated LTI: asymptotic stability implies global
eigenvalues on the imaginary axis (unstable). exponential stability.
11 12
State Trajectories Example
3
Is the system asymptotically stable?
2
Stable
>> A=[3,-2,0;8,-3,-4;0,4,-9];
1
>> eig(A)
ans =
0

-1.0000
Asymptotically stable
-1
-3.0000
-2 -5.0000
Yes. All poles in the open LHP.
-1.5 -1 -0.5 0 0.5 1

13 14

Example Input-output (BIBO) Stability


Is the system stable? Definition For any bounded input, the output is
>> A=[-5,5,-5;-20,20,-20;-20,20,-20]; bounded.
>> eig(A)
ans = Definition can be generalized to time-varying
0 MIMO systems using ||.|| (norm) in place of |.|
-5 Definition can be generalized to distributed
0 parameter systems.
No. Repeated pole on the imaginary axis.
15 16
Contradiction
Theorem 1: BIBO Stability Proof Theorem 1 BIBO Stability
A SISO LTI system is BIBO stable if and only Sufficiency (if)
if its impulse response satisfies

Remarks Necessity (only if): Assume BIBO stable with condition


Condition can be generalized to time-varying violated and let
MIMO systems using ||.|| (norm) in place of |.|
Condition can be generalized to distributed
parameter systems.
17 18

Theorem 2: BIBO Stability Proof: Sufficiency


,
LTI SISO system is BIBO stable if and only if all its
transfer function poles are in the open LFP.
Proof (Necessity)

After pole-zero cancellation, poles =remaining


The integral of diverges for any transfer LHP poles (not all ) BIBO stability.
function pole is in the closed RHP . 19 20
Relationship Between Internal
Kalman Decomposition Stability BIBO Stability
Y(s)
U(s) Controllable BIBO stability is equivalent to open LHP poles
Any system can be Observable
Internal stability implies BIBO stability (since
decomposed into four Y(s) poles are a subset of the eigenvalues).
Uncontrollable
subsystems as shown Observable
Some eigenvalues may cancel in the transfer
in the figure:
Us) function and are not poles.
Unobservable Mode Controllable
Unobservable BIBO stability does not , in general, imply internal
stability
Uncontrollable Mode Uncontrollable
Unobservable With no cancellation, {poles}={eigenvalues}
BIBO stability is equivalent to internal stability
21 22

Important Relations
Definitions
x2 Internally stable systems are stabilizable
Observable
Detectable: all unstable Subspace and detectable (no unstable modes).
modes are observable Unobservable
Subspace
y Observable systems are detectable (no
(i.e. all unobservable (stable)
x1 unstable unobservable modes).
modes are stable).
Controllable systems are stabilizable (no
Stabilizable: all unstable x2 Controllable unstable uncontrollable modes).
modes are controllable Uncontrollable Subspace

(i.e. all uncontrollable Subspace For minimal realizations, BIBO stability


(stable) u and internal stability are equivalent
modes are stable). x1
{poles}={eigenvalues}.
23 24
Example Example (continued)
s 1 s 1 1 s 1 s 1 1
G( s) 2 G( s) 2
s 1s 1 s 1 s 1 s 1s 1 s 1 s 1
1 0 BIBO stable 1 0 BIBO stable
Controllable Form : without cancellation Observable Form : without cancellation
0 1 0 1 1 0 1 1 1 1
A B C 1 1 O A B C 1 0 C
1 0 1 1 1
1 0 1 1 1

Controllable (phase var. form) but not observable. Observable (observer form) but not controllable.
BIBO stable but not internally stable. BIBO stable but not internally stable.
Stabilizable but not detectable. Detectable but not stabilizable.
25 26