01.05.2017 Views

563489578934

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

82<br />

Signals and Spectra Chap. 2<br />

Input waveform<br />

x(t)<br />

Linear system<br />

Output waveform<br />

y(t)<br />

h(t)<br />

H(f)<br />

Some descriptions<br />

for the input<br />

X(f)<br />

R x ()<br />

“Voltage” spectrum<br />

Autocorrelation function<br />

Power spectral density<br />

Figure 2–14 Linear system.<br />

Some descriptions<br />

for the output<br />

Y(f)<br />

R y ()<br />

2–6 REVIEW OF LINEAR SYSTEMS<br />

Linear Time-Invariant Systems<br />

An electronic filter or system is linear when superposition holds—that is, when<br />

y1t2 = [a 1 x 1 1t2 + a 2 x 2 1t2] = a 1 [x 1 1t2] + a 2 [x 2 1t2] (2–130)<br />

where y(t) is the output and x(t) = a1 x 1 (t) + a 2 x 2 (t) is the input, as shown in Fig. 2–14. [ # ]<br />

denotes the linear (differential equation) system operator acting on [ # ] . The system is said to<br />

be time invariant if, for any delayed input x(t - t 0 ), the output is delayed by just the same<br />

amount y(t - t 0 ). That is, the shape of the response is the same no matter when the input is applied<br />

to the system.<br />

A detailed discussion of the theory and practice of linear systems is beyond the scope of<br />

this book. That would require a book in itself [Irwin, 1995, 2011]. However, some basic ideas<br />

that are especially relevant to communication problems will be reviewed here.<br />

Impulse Response<br />

The linear time-invariant system without delay blocks is described by a linear ordinary differential<br />

equation with constant coefficients and may be characterized by its impulse response<br />

h(t). The impulse response is the solution to the differential equation when the forcing function<br />

is a Dirac delta function. That is, y(t) = h(t) when x(t) = d(t) . In physical networks, the<br />

impulse response has to be causal. That is, h(t) = 0 for t 6 0. † This impulse response may<br />

be used to obtain the system output when the input is not an impulse. In that case, a general<br />

waveform at the input may be approximated by ‡<br />

† The Paley–Wiener criterion gives the frequency domain equivalence for the causality condition of the time<br />

domain. It is that H( f) must satisfy the condition<br />

‡ ∆t corresponds to dx of Eq. (2–47).<br />

q<br />

L-q<br />

| ln |H(f)| |<br />

1 + f 2 df 6q

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!