Scribd
Upload
48 views3 pages

ECE 562: Problem Set 6 Linear Equalizers, Signaling On Fading Channels

Uploaded by

Debajyoti Datta
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
48 views3 pages

ECE 562: Problem Set 6 Linear Equalizers, Signaling On Fading Channels

Uploaded by

Debajyoti Datta
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

University of Illinois Fall 2016

ECE 562: Problem Set 6


Linear Equalizers, Signaling on Fading Channels
Due: Thursday, November 17 in class (only turn in solutions to Problems 1-7)
Reading: Lecture Notes 16-19, channel modeling slides

1. [MSE of linear equalizers]


The MSE for the k-th symbol when a linear equalizer ck is used was defined in class to be:
h i
MSEk = E |c†k Z − smk |2

Using the fact that X


Z = hk smk + hj smj + W
j6=k

show that
MSEk = Es |c†k hk − 1|2 + Es |c†k hj |2 + N0 kck k2
X

j6=k

2. [Minimum MSE]
Consider the MMSE equalizer ck,MMSE = Es (Es H H † + N0 I)−1 hk . Show that MSE achieved by this
equalizer (which is the minimum MSE) is given by:

MMSEk = Es − Es2 h†k (Es H H † + N0 I)−1 hk

3. [MMSE equallizer as SNR goes to infinity]


As we discussed in class, the MMSE equalizer becomes ill-conditioned if we set N0 = 0. Interestingly,
however, we can still show that the MMSE equalizer converges to the ZF equalizer in limit as N0 → 0.
(a) Show that MSEk for ck,ZF is given by:

MSEk (ck,ZF ) = N0 kck,ZF k2 .

(b) Using part (a) argue that, for a fixed Es , MSEk for ck,MMSE satisfies:

lim MSEk (ck,MMSE ) = 0.


N0 →0

(c) Now use the results of part (b) and Problem 2 to conclude that the MMSE equalizer indeed
converges to the ZF solution as N0 → 0, for fixed Es .
4. [MMSE equalizer maximizes SINR]
We showed in class that the SINR for the k-th symbol when a linear equalizer ck is used is given by:

Es |c†k hk |2
SINRk =
|c†k hj |2 + N0 kck k2
P
Es j6=k

Using the result of Problem 2, show that ck,MMSE maximizes SINRk .


Hint: Consider the problem of maximizing SINRk subject to c†k hk = α, and show that the achieved
maximum is independent of α.
5. [Maximum SINR and Minimum MSE]
Show that the maximum value of SINRk is given by
Es
SINRk,max = −1
MMSEk
6. [True or False]
Determine if the following statements are True or False. You need to provide a brief justification for
your answer to get credit.
(a) If the mobile speed is 72 km/hr, the carrier frequency is 900 MHz, and symbol rate for commu-
nication is 10 symbols a second, then the mobile experiences slow fading.
(b) In wireless communication, if the speed of the mobile is doubled the delay spread is also doubled.
(c) A wireless channel with τds = 10−4 seconds and fmax = 100 Hz operating with a bandwidth of
1.25 MHz is a flat fading channel.
(d) For a fixed γ b , Pb for binary DPSK signaling over a slow flat fading channel is always larger than
Pb without fading, irrespective of the distribution of the fading.
7. [Useful Result]
Show that Z ∞ √ e−x/γ
 r 
1 γ
Q( x) dx = 1− .
0 γ 2 2+γ
Hint: Start with:
Z ∞ Z ∞ Z ∞ 2 Z ∞ Z t2 −t2 /2 −x/γ
√ e−x/γ e−t /2 e−x/γ e e
Q( x) dx = √
dt dx = dx dt
x=0 γ x=0 t= x 2π γ t=0 x=0 2π γ
and use the fact that the Gaussian pdf integrates to 1 to conclude the result.

You do not need to turn in the solutions to Problems 8-10. I’ll give you the solutions to
these next Thursday when you turn in the solutions to Problems 1-7.

8. [MPSK Signaling in Rayleigh Fading]


For MPSK signaling,
1 (M −1)π/M γs sin2 (π/M )
Z  
Pe (γs ) = exp − dθ .
π 0 sin2 θ
Using this expression how that the average symbol error probability Pe for MPSK signaling in Rayleigh
fading is given in closed form by
    
1 1 1 1 cot π/M
Pe = 1 − −√ + tan−1 √ ,
M 1 + a2 2 π 1 + a2
1
where a2 = γ s sin2 π/M
.
Hint: You may need to use the following integral
Z θ2
1 1

1

cotθ
 
π θ2
−1
2 2
dθ = √ tan √ − − θ for 0 ≤ θ1 ≤ θ2 ≤ π/2.
θ1 cosec θ + a a2 1 + a2 1 + a2 2 θ1

9. [Diversity]
Consider BPSK with channel gain a, i.e., the received signal is
√
r(t) = ±a Eg(t) + w(t), 0 ≤ t ≤ T,
where {w(t)} is a zero-mean complex WGN process with PSD N0 , g(t) is a unit energy signal, and the
channel gain a is random with probability mass function
P{a = 0} = 0.1 and P{a = 2} = 0.9.
2
(a) Determine the average probability of error Pe for MPE detection.
(b) What value does Pe approach as E/N0 approaches infinity?
(c) Suppose the same signal is transmitted on two statistically independent channels with gains a1
and a2 , where

P{a1 = 0} = P{a2 = 0} = 0.1 and P{a1 = 2} = P{a2 = 2} = 0.9.

The additive noises on the two channels are also independent and identically distributed. The
demodulator employs a matched filter for each channel and adds the two filter outputs to form
the decision variable (which is compared to 0 for decision-making). Determine Pe in this case.
(d) For the case in part (c), what value does Pe approach when E/N0 approaches infinity?

10. [Optimality of maximal-ratio combining]


Consider BPSK signaling on an L-th order diversity channel. Each channel introduces a fixed attenu-
ation and phase shift so that the received signal at the output of the `-th channel is:
√
r` (t) = ±α` ejφ` E g` (t) + w` (t)

where the processes w` (t) are independent complex WGN processes with PSD N0 .
The receiver uses the decision statistic
L
X
R= β` hr` , g` i
`=1

where the {β` } are complex weighting factors to be determined. A decision in favor of +1 (“symbol
0”) is made if rI > 0 and −1 (“symbol 0”) otherwise.
(a) Determine the p.d.f. of RI when +1 is transmitted.
(b) Show that the probability of bit error Pb is given by:
r 
PL jφ`
2E `=1 Re{β` α` e }
Pb = Q  qP .
N0 L 2
`=1 |β ` |

(c) Determine the values of {β` } that minimize Pb .


Hint: Use the Cauchy-Schwarz inequality

You might also like