Multipath Propagation Rev6

Download as pdf or txt
Download as pdf or txt
You are on page 1of 1

Multipath Propagation

I.

Analytic
Look over the solution to Problem #8.64 in [1].

II.

Numerical
Fix n0 = 4 and use the same vertical axis dynamic range on all FFT plots since the desire is to
compare them. Use (radians/sample) or f (cycles/sample) as the frequency axis rather than k
(FFT bin index). There is no need to plot the negative frequencies.
Note: H1 (k ) =

1
1
samples the Fourier transform of the true inverse system Hi (z ) =
H( k )
H (z )

A.

Plot h(n) and the zero locations of H(z). Augment h(n) with zeros out to NFFT = 256,
FFT, and plot linear magnitude and phase of H(k) vs. f to illustrate the amplitude and
phase distortion caused by the multipath.

B.

Let N = 16
1.

Obtain the N-point FFT H(k) of h(n) and the N-point FFT H1(k) of h1(n). Plot
linear magnitude and phase of H(k) and H1(k) vs. f.

2.

Obtain h1(n) and plot h1(n) and the zero locations of H1(z).

3.

Augment h1(n) with zeros out to NFFT = 256, FFT, and plot linear magnitude
and phase of H1(k) vs. f.

4.

Obtain the linear convolution h2(n) = h(n) * h1(n) via FFT (2N = 32 point
result).

5.

Plot h2(n) and zero locations of H2(z). Note: Trailing zeros in the 2N-point
result for h2(n) should be truncated prior to determining zero locations for
H2(z).

6.

C.

Augment h2(n) with zeros out to NFFT = 256, FFT, and plot linear magnitude
and phase of H2(k) vs. f to see how well h1(n) has equalized the channel.

Let N = 64
Repeat II B [1 - 6]. Note that h1(n) closely resembles hi(n) out through n = 63. Part (4)
yields a 2N = 128 point result. In C6, add a pair of plots that blows up the region near
|H2(k)| =1 and arg{H2(k)| = 0 to show the residual distortion.

References
[1]

A. Oppenheim, R. Schafer, and J. Buck. Discrete-Time Signal Processing. 2nd Ed. Prentice-Hall
(1999). Note same as Problem #8.67 in Discrete-Time Signal Processing 3rd Ed. (2010).

You might also like