Digital Signal Processing Reference
In-Depth Information
In this section, we derive an equivalent condition to determine the stability of
an LTIC system modeled with a rational Laplace transfer function
H
(
s
)given
in Eq. (6.35). Since we are mostly interested in causal systems, we assume
that the Laplace transfer function
H
(
s
) corresponds to a right-sided system.
The poles of a system with a transfer function as given in Eq. (6.35) can be
calculated by solving the characteristic equation, Eq. (6.36). Three types of
poles are possible. Out of the
n
possible poles, assume that there are
L
poles at
s
=
0,
K
real poles at
s
=−σ
k
,1
≤
k
≤
K
, and
M
pairs of complex-conjugate
poles at
s
=−α
m
j
ω
m
,1
≤
m
≤
M
, such that
L
+
K
+
2
M
=
n
. In terms
of its poles, the transfer function, Eq. (6.35), is given by
H
(
s
)
=
N
(
s
)
D
(
s
)
N
(
s
)
=
.
(6.40)
K
M
s
2
+
2
α
m
s
+
α
m
+ ω
m
s
L
(
s
+ σ
k
)
k
=
1
m
=
1
From Table 6.1, the repeated roots at
s
=
0 correspond to the following term
in the time domain:
1
n
!
t
n
u
(
t
)
1
s
n
.
←→
(6.41)
Since term
t
n
u
(
t
) is unbounded as
t
→∞
, a stable LTIC system will not contain
such unstable terms. Therefore, we assume that
L
=
0. The partial fraction
expansion of Eq. (6.40) with
L
=
0 results in the following expression:
A
1
(
s
+ σ
1
)
A
K
(
s
+ σ
K
)
B
1
s
+
C
1
H
(
s
)
=
++
+
+
α
1
+ ω
1
s
2
+
2
α
1
s
+
B
M
s
+
C
M
+
,
(6.42)
s
2
+
2
α
M
s
+
α
2
M
+ ω
2
M
where
{
A
k
,
B
m
,
C
m
}
are the partial fraction coefficients. Calculating the inverse
Laplace transform of Eq. (6.42) and assuming a causal system, we obtain the
following expression for the impulse response
h
(
t
) of the LTIC system:
K
M
−σ
k
t
u
(
t
)
−α
m
t
cos(
ω
m
t
+
θ
m
)
u
(
t
)
h
(
t
)
=
A
k
e
+
r
m
e
,
(6.43)
k
=
1
m
=
1
h
k
(
t
)
h
m
(
t
)
where we have expressed the terms with conjugate poles in the polar format.
Constants
{
r
m
,
θ
m
}
are determined from the values of the partial fraction coef-
ficients
{
B
m
,
C
m
}
and
α
m
.
In Eq. (6.43), we have two types of terms on the right-hand side of the
equation. Summation I consists of
K
real exponential functions of the type
h
k
(
t
)
=
A
k
exp(
−σ
k
t
)
u
(
t
). Depending upon the value of
σ
k
, each of these func-
tions
h
k
(
t
) may have a constant, decaying exponential or a rising exponential
waveform.
Summation II consists of exponentially modulated sinusoidal functions of
the type
h
m
(
t
)
=
r
m
exp(
−α
m
t
) cos(
ω
m
t
+ θ
m
)
u
(
t
). The stability characteristic
Search WWH ::
Custom Search