Electrical Engineering Department

University of Hail
Hail-2019

Advanced Digital Control Syst

EE554
Chapter 3
Stability of digital control systems

Dr Mourad Kchaou
Electrical Engineering Department
Hail University

Friday 13th March, 2020

Presented by Dr Mourad v. (b.2003131104)

Outline

1 Asymptotic stability

2 BIBO Stability

3 Jury’s stability Test

4 Lyapunov stability theory

Asymptotic stability

Asymptotic stability

Asymptotic stability
We will discuss in this chapter, the asymptotic stability of the
state space representation and the
bounded-input-bounded-output (BIBO) stability of the impulse
transfer functions.
The natural response of a linear system depends on the initial
condition. It may:
1. asymptotically stable it converges to the origin
asymptotically
2. marginally stable it remains in a bounded region in the
vicinity of the origin,
3. unstable it grows unbounded.

Introduction 3 / 1
Asymptotic stability

Equilibrium point

Consider a nonlinear system described as follows

(
x(k + 1) = f (x(k))
(1)
x(0) = x0 6= 0

Equilibrium point
xe ∈ Rn is called an equilibrium point of (1), if

f (xe ) = xe

For linear state equations, we have Axe = xe . Thus, a linear

system has one equilibrium state at the origin.
Introduction 4 / 1
Asymptotic stability

Equilibrium point
Find the equilibrium points of the following systems:
Example

x(k + 1) = x(k)(2x(k) − 0.1) (2)

At equilibrium, we have

xe = xe (2xe − 0.2) (3)

We rewrite the equilibrium condition as

xe (2xe − 1.2) = 0 (4)

Hence, the system has the two equilibrium states xe = 0 and

xe = 0.6
Introduction 5 / 1
Asymptotic stability

Asymptotic Stability

Equilibrium point
A linear system is said to be asymptotically stable, if all its
trajectories converge to the origin-that is, for any initial state
x(k0 ), x(k) −→ 0 as k −→ ∞.

Introduction 6 / 1
BIBO Stability

BIBO Stability

Theorem
A discrete-time linear system is asymptotically stable if and
only if all the eigenvalues of its state matrix are inside the
unit circle.

Introduction 7 / 1
BIBO Stability

BIBO Stability

Theorem
A system is BIBO stable if its output is bounded for any
bounded input. That is,

ku(k)k < α < ∞)resultsaboundedoutput(ky(k)k < β < ∞ (5)

Theorem
A system is BIBO stable if and only if the norm of its impulse
response matrix is absolutely summable. That is,
∞
X
kH(k)k < ∞ (6)
k=0

Introduction 8 / 1
BIBO Stability

BIBO Stability
Determine the BIBO stability of the system with difference
equations

y1 (k + 2) + 0.1y2 (k + 1) = u(k) (7)

y2 (k + 2) + 0.9y1 (k + 1) = u(k) (8)

Using the z transform, we find

 2 
z 0.1z
Y (z) 2 = U (z) (9)
z 0.9z

Thus
 
z −0.1
−1
z 2 0.1z −0.9 z

G(z) = = (10)
0.9z z 2 2
z(z − 0.09)

Introduction 9 / 1
BIBO Stability

BIBO Stability

Inverse z-transforming the matrix gives the impulse response

" #
5.556((−0.3)k + (0.3)k ) 1.111δ(k − 1) + 1.852((−0.3)k − (0.3)k )
G(k) =
10δ(k − 1) + 16.667((−0.3)k − (0.3)k ) k k
5.556((−0.3) − (0.3) )
(11)

The entries of the impulse response matrix are all absolutely summable. Hence, the system is
BIBO stable.

Introduction 10 / 1
Jury’s stability Test

Jury’s stability Test

To investigate the stability of z-domain polynomials, with real

coefficients, the Jury test can be performed.
Consider the transfer function G(z) for a linear discrete-time
system.

N (z)
G(z) =
D(z)

D(z) = a0 + a1 z + an−2 z n−2 + · · · + an−1 z n−1 + an z n = 0

with an > 0 and ai réels.

Introduction 11 / 1
Jury’s stability Test

Jury’s stability Test

We construct the following table:

1 a0 a1 a2 ··· ··· an−1 an

2 an an−1 an−2 ··· ··· a1 a0
3 b0 b1 b2 ··· ··· bn−1
4 bn−1 bn−2 bn−3 ··· ··· b0
5 c0 c1 c2 ··· cn−2
6 cn−2 cn−3 cn−4 ··· c0
.. .. .. ..
. . . .
2n-5 p0 p1 p2 p3
2n-4 p3 p2 p1 p0
2n-3 q0 q1 q2

Introduction 12 / 1
Jury’s stability Test

Jury’s stability Test

where
 
 a0 an−k 
bk =   , k = 0, 1, · · · n − 1
an ak 
 
 b0 bn−1−k 
ck = 
  , k = 0, 1, · · · n − 2
bn−1 bk 
 
p0 p3−k 
qk = 
 
p3 pk 

Introduction 13 / 1
Jury’s stability Test

Jury’s stability Test

Jury’s stability Test

the roots of the polynomial are inside the unit circle if and only
if
1. D(1) > 0
2. (−1)n D(−1) > 0


 |a0 | < an

|b0 | > |bn−1 |



3. |c0 | > |cn−2 |

|p0 | > |p3 |





|q0 | > |q2 |


Introduction 14 / 1
Jury’s stability Test

Jury’s stability Test

Cas n = 1 then, D(z) = a0 + a1 z Jury conditions
1. a0 + a1 > 0;
2. −a0 + a1 > 0;
3. |a0 | < a1 .
Cas n = 2 then, D(z) = a0 + a1 z + a2 z 2 Jury conditions
1. a0 + a1 + a2 > 0;
2. a0 − a1 + a2 > 0;
3. |a0 | < a2 .
Cas n = 3 then, D(z) = a0 + a1 z + a2 z 2 + a3 z 3 Jury conditions
1. a0 + a1 + a2 + a3 > 0;
−a0 + a1 − a2 + a3 > 0;
2. 
|a0 | < a3

3. |b0 | > |b2 |

|c0 | > |c1 |


Introduction 15 / 1
Jury’s stability Test

Jury’s stability Test

Consider the transfer function:

z
G(z) = 4
3z + 2 + z 2 + z + 1
3

the characteristic polynomial is

D(z) = 3z 4 + 23 + z 2 + z + 1 = 0

we have
D(1) = 8 > 0
(−1)4 D(−1) = 2 > 0

Introduction 16 / 1
Jury’s stability Test

Jury’s stability Test

r1 1 1 1 2 3
r2 3 2 1 1 1
r3 -8 -5 -2 -1
r4 -1 -2 -5 -8
r5 63 38 11

|a0 | = 1 < 3 = a4
|b0 | = 8 > 1 = |b3 |
|c0 | = 63 > 11 = |c2 |

The system is stable.

Introduction 17 / 1
Lyapunov stability theory

Lyapunov stability theory

Lyapunov stability theory is based on the idea that at a stable

equilibrium, the energy of the system has a local minimum,
whereas at an unstable equilibrium, it is at a maximum.
This property is not restricted to energy and is in fact shared
by a class of function that depends on the dynamics of the
system. We call such functions Lyapunov functions.

Introduction 18 / 1
Lyapunov stability theory

Lyapunov functions
If a Lyapunov function can be found for an equilibrium point,
then it can be used to determine its stability or instability.
We begin by examining the properties of energy functions that
we need to generalize and retain for a Lyapunov function. We
note the following:
I Energy is a nonnegative quantity.
I Energy changes continuously with its arguments
.

A scalar continuous function V (x) is said to be positive definite

if
I V (0) = 0.
I V (x) > 0 for any non-zero x.

Introduction 19 / 1
Lyapunov stability theory

Lyapunov stability theory

Lyapunov Function
A scalar function V (x) is a Lyapunov function in a region D if
it satisfies the following conditions in D:
I It is positive definite.
I It decreases along the trajectories of the system-that is,
∆V (k) = V (x(k + 1)) − V (x(k)) < 0

Introduction 20 / 1
Lyapunov stability theory

Lyapunov stability theory

Consider the system

x(k + 1) = Ax(k) (12)

Theorem
System (12) is stable if and only if given any positive definite
(PD), Q, the solution P of the equation:

P = AP A + Q (13)

Introduction 21 / 1