This preview shows pages 1–2. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: Sequences with Rational Power Spectra
Consider the class of zeromean, ﬁnite variance, WSS, random sequences with
rational PSD. These random signals are in general obtained by the processing of
a zeromean, WSS, white, random signal through a LTI system and consequently
the general form for the PSD of these signals is:
2
Pxx (z ) = σo B (z )B ∗
A(z )A∗ 1
z∗
1
z∗ If the PSD of the signal contains just P poles, i.e, there are no zeroes or in other
words B (z ) = K then the PSD takes the form:
Pxx (z ) = 2
Kσo
∗
A(z )A 1
z∗ These processes are termed as autoregressive or AR(p) random signals. It is not
diﬃcult to show that the ACF of the random sequence x[n] in this case takes
the form:
p k  Rxx [k ] = ci ai , ai  < 1
i Note that Gaussian AR random signals are ergodic in the general sense because:
∞ Rxx [k ] < ∞.
k=−∞ If the PSD of the random signal has just q zeroes in it, i.e, A(z ) = 1 then the
PSD takes the form:
1
2
Pxx (z ) = σo B (z )B ∗
.
z∗
These processes are termed as moving average or MA(q ) random signals and
the ACF of the random sequence can be shown to satisfy:
Rxx [k ] = 0, k  > q Signals with both poles and zeroes in their PSD are referred to as autoregressive
moving average or ARMA(p, q ) random signals. 2 AR PROCESS, p = 1, L = 2000, σo = 0.5
SIGNAL 0.5 0 −0.5 SAMPLE REALIZATION OF 1 AR(1) 1.5 −1 −1.5
0 500 1000
TIME SAMPLES 1500 2000 (a) 2
AR PROCESS, p = 1, L = 2000, σo = 0.5 1.2
1 Rxx[k] 0.8
0.6
0.4
0.2
0
−0.2
−20 −15 −10 −5
0
5
10
ACF LAG VARIABLE (k) 15 20 (b) 2 AR PROCESS, p = 1, L = 2000, σo = 0.5
1.4 ACT
THEO 1.2
1 Pxx(ej ω) 0.8
0.6
0.4
0.2
0
0 0.1 0.2
0.3
NORMALIZED FREQUENCY 0.4 0.5 (c) Figure 1: AR(1) example: (a) sample realization of a AR(1) process, (b) sample
ACF of the process, and (c) estimated PSD of the process. ...
View Full
Document
 Fall '11
 GraceGraham

Click to edit the document details