combat fading in wireless

How to combat fading in wireless channels?

1. A. Chandra, NIT Durgapur – How to combat fading?
1
J.U., 13
th
April, 2007
How to Combat Fading in How to Combat Fading in
Wireless Channels?Wireless Channels?
Aniruddha Chandra
ECE Department, NIT Durgapur
[email protected]

2. A. Chandra, NIT Durgapur – How to combat fading?
2
J.U., 13
th
April, 2007
Fading in Wireless Channels
Mitigation of Slow Flat Fading
Mitigation of Frequency Selective Fading
Mitigation of Fast Fading
Outline

3. A. Chandra, NIT Durgapur – How to combat fading?
3
J.U., 13
th
April, 2007
Fading in Wireless Channels
Mitigation of Slow Flat Fading
Mitigation of Frequency Selective Fading
Mitigation of Fast Fading
What is Fading?
Fading Mechanisms
Degradation due to Fading

4. A. Chandra, NIT Durgapur – How to combat fading?
4
J.U., 13
th
April, 2007
Why Wireless?
•Mobility Anytime, Anywhere connectivity
Phone for people not for places
•Easy Installation Rapid deployment, reconfigurable, no cable,
easy maintenance
•Digital Companion Voice, message, internet, multimedia
Hi! I’m on
the prowl

5. A. Chandra, NIT Durgapur – How to combat fading?
5
J.U., 13
th
April, 2007
Fading???
•Fading over time
•Fading over distance
Enemy
ahead
Enemy
?
Oh No ..
Emily!
Boss’s
wife
Lunch
break, at
last!

6. A. Chandra, NIT Durgapur – How to combat fading?
6
J.U., 13
th
April, 2007
Radio Wave Propagation
Street Sign
STOP
Line of Sight
Reflection
Diffraction
Scattering
Transmitter
Receiver
Buildings
Earth surface
•Multiple replica of signal combines with random phase resulting in random
amplitude attenuation and phase variation  This is Fading
•Relative velocity causes further change with time.
Fading
Envelope
Distance
A
m
p
l
i
t
u
d
e
q
l
=cos
v
f
D

7. A. Chandra, NIT Durgapur – How to combat fading?
7
J.U., 13
th
April, 2007
Multipath
t
0
+τ
1
t
0
t
0
+τ
2
Time Variance
•Limited Power
Size, Weight, Battery Constraints
•Limited BW
Spectrum allocation
•Path Loss
up to 10 dB/km
•Multipath Fading
Resolvable channel induced ISI
Non-resolvable  random amplitude variation
•Time Variance of the Channel
Due to relative velocity
Introduces Doppler Effect
Wireless Channel

8. A. Chandra, NIT Durgapur – How to combat fading?
8
J.U., 13
th
April, 2007
Fading
Large Scale
Fading
Small Scale
Fading
Attenuation
with distance
Variation
about mean
Time delay
spread /Multipath
Doppler shift /
time variance of
channel
Flat Fading
Frequency
Selective Fading
Fast
Fading
Slow
Fading
Fading Mechanisms
•Empirical Model
•Okumura, Hata, COST 231
•Statistical Model
•Rayleigh, Rice, Nakagami-m, Hoyt, Log-Normal, Weibull, Gamma, K etc.

9. A. Chandra, NIT Durgapur – How to combat fading?
9
J.U., 13
th
April, 2007
Fading
Large Scale
Fading
Small Scale
Fading
Large Scale & Small Scale Fading
•Large Scale Fading
•Small Scale Fading
•Due to general terrain, density and height of buildings, vegetation
•Variation occurs over very large distances (100m.-a few K.m.)
•Important for predicting the coverage and availability of a particular service
•Due to local environment, nearby trees, buildings
•Variation occurs over very short distances, on the order of the signal
wavelength (<1 m.)
•Important for design of modulation format and transmitter / receiver design

10. A. Chandra, NIT Durgapur – How to combat fading?
10
J.U., 13
th
April, 2007
Fading
Large Scale
Fading
Attenuation
with distance
Variation
about mean
Large Scale Fading
•Attenuation with Distance / Path Loss
•Variation about Mean / Shadowing
•Friis equation for free space

•Empirical models (Okumura, Hata etc.) based on field measurements
•Characterized by log-normal shadowing
•If the distribution of is log-normal with parameters and
2
4
÷
ø
ö
ç
è
æ
p
l
µ
dP
P
t
r
Path Loss alone
Shadowing and
Path Loss
log (d)
P
r
/P
t

(in dB)
Slope 10n dB/ decade
Generally n>>2
(n=2 for free space)
()
( )
ú
ú
û
ù
ê
ê
ë
é
s
m-y
-
psy
x
=y
y
y
y
2
2
10
2 2
log10
exp
2
f
rtPP=y y
y
m
y
s
10ln10=xwhere

11. A. Chandra, NIT Durgapur – How to combat fading?
11
J.U., 13
th
April, 2007
Fading
Small Scale
Fading
Time delay
spread /Multipath
Doppler shift /
time variance of
channel
Flat Fading
Frequency
Selective Fading
Fast
Fading
Slow
Fading
Small Scale Fading
•Slow Flat Fading
•Least severe fading type
•Multiplicative narrowband fading  Modeled with statistical distributions

12. A. Chandra, NIT Durgapur – How to combat fading?
12
J.U., 13
th
April, 2007
()
÷
÷
ø
ö
ç
ç
è
æ
s
a
-
s
a
=a
2
2
2
2
expf
() ÷
ø
ö
ç
è
æ
s
a
÷
÷
ø
ö
ç
ç
è
æ
s
+a
-
s
a
=a
202
22
2
2
exp
s
I
s
f
2
2
2s
=
s
K
()
()
÷
ø
ö
ç
è
æ
a
W
-a÷
ø
ö
ç
è
æ
WG
=a
- 212
exp
2 mm
m
f
m
m
2
a=W
•Rayleigh
•Rician
•Nakagami-m
Small Scale Fading (Contd.)

13. A. Chandra, NIT Durgapur – How to combat fading?
13
J.U., 13
th
April, 2007
Fading
Small Scale
Fading
Time delay
spread /Multipath
Doppler shift /
time variance of
channel
Frequency
Selective Fading
Slow
Fading
Ground to Ground: Highly frequency
selective, not very time selective
Small Scale Fading (Contd.)

14. A. Chandra, NIT Durgapur – How to combat fading?
14
J.U., 13
th
April, 2007
Fading
Small Scale
Fading
Time delay
spread /Multipath
Doppler shift /
time variance of
channel
Flat Fading
Fast
Fading
Air to Air: Almost frequency non-
selective, very time selective
Small Scale Fading (Contd.)

15. A. Chandra, NIT Durgapur – How to combat fading?
15
J.U., 13
th
April, 2007
Fading
Small Scale
Fading
Time delay
spread /Multipath
Doppler shift /
time variance of
channel
Frequency
Selective Fading
Fast
Fading
Air to Ground: Frequency selective,
time selective
Small Scale Fading (Contd.)

16. A. Chandra, NIT Durgapur – How to combat fading?
16
J.U., 13
th
April, 2007
Degradation Due to Fading
AWGN
ò
T
0
dt
sin (ω
C
t) sin (ω
C
t)
Channel ReceiverTransmitter
Binary baseband digital data
ASK modulated waveform Modulated signal + Noise
0 1 1 1 1 0 0 1 0 0
Demodulated signal
Error
0 0 1 1 1 0 0 1 0 0

17. A. Chandra, NIT Durgapur – How to combat fading?
17
J.U., 13
th
April, 2007
Degradation Due to Fading (Contd.)
AWGN
ò
T
0
dt
sin (ω
C
t) sin (ω
C
t)
Rayleigh
Fading
Channel ReceiverTransmitter
Binary baseband digital data
ASK modulated waveform
Modulated signal + Noise
Rayleigh faded signal
0 1 1 1 1 0 0 1 0 0 0 0 0 1 1 1 0 1 0 0
Demodulated signal
Too Many Errors

18. A. Chandra, NIT Durgapur – How to combat fading?
18
J.U., 13
th
April, 2007
The Good, the Bad & the Ugly!
B. Sklar, ‘Rayleigh fading channels in digital
communication systems’, IEEE Communications
Magazine, July, 1997.
•Good (AWGN channel)
Exponential decrease
•Bad (Slow, flat fading channel)
Linear decrease
Loss in SNR
•Ugly (Frequency selective/
fast fading channel)
Irreducible error floor
?
Loss

19. A. Chandra, NIT Durgapur – How to combat fading?
19
J.U., 13
th
April, 2007
Fading in Wireless Channels
Mitigation of Slow Flat Fading
Mitigation of Frequency Selective Fading
Mitigation of Fast Fading
What is Diversity?
Diversity Types
Diversity Combining
Error Correction Coding

20. A. Chandra, NIT Durgapur – How to combat fading?
20
J.U., 13
th
April, 2007
Improvement with Diversity
ò
T
0
dt
sin (ω
C
t)
Diversity
Combiner
AWGN
Rayleigh
Fading
AWGN
Rayleigh
Fading
Channel Receiver
Rayleigh faded signal # path1
Rayleigh faded signal # path2
Output of Diversity Combiner
0 0 1 1 1 1 0 1 0 0
Lesser Errors
Demodulated signal

21. A. Chandra, NIT Durgapur – How to combat fading?
21
J.U., 13
th
April, 2007
Why Diversity?
BER performance of BPSK with Rayleigh fading and
subsequent improvement with 2
nd
order diversity
Gain
•Mitigate slow flat fading?
Increase the transmitted power
Not power efficient technique
•Alternative way 
Diversity•If one signal path undergoes a deep
fade at a particular point of time,
another independent path may have a
strong signal
•If probability of a deep fade in one
channel is p, then the probability for L
channels is p
L

22. A. Chandra, NIT Durgapur – How to combat fading?
22
J.U., 13
th
April, 2007
Macroscopic & Microscopic Diversity
Base
Station
A
Base
Station
B
Hill
Hill
Reception
from only A
Reception
from only B
Shadowed
Region
•Macroscopic Diversity
Mitigate effects of large scale fading
or, shadowing
By selecting a base station which is
not shadowed when others are
•Microscopic Diversity
Mitigate effects of small scale fading or, multipath
Require two or more uncorrelated received signals, with the same long-
term fading experienced in those signals

23. A. Chandra, NIT Durgapur – How to combat fading?
23
J.U., 13
th
April, 2007
Diversity Types
Receiver
CombinerTransmitter
Receiver •Receiver Diversity (SIMO)
Antenna separation λ/2
•Transmit Diversity (MISO)
Antenna separation 10λ
The total transmitted power is
split among the antennas
Open loop/ close loop (for 3G)
Transmitter
Transmitter
Receiver Combiner
•Space Diversity
•Spatial separation between antennas, so that the diversity branches
experience uncorrelated fading
•More hardware/ antennas

24. A. Chandra, NIT Durgapur – How to combat fading?
24
J.U., 13
th
April, 2007
Diversity Types (Contd.)
ReceiverTransmitter
Path 2
Path 1
freq
time
f
1 (path 1)
f
2 (path 2)
•Frequency Diversity
•Modulate the signal through L different carriers
•The separation between the carriers should be at least the coherent bandwidth,
not effective over frequency-flat channel
•Only one antenna is needed
•The total transmitted power is split among the carriers, not BW efficient
ReceiverTransmitter
Path 2
Path 1
freq
time
t
1
(path 1)
t
2
(path 2)
•Time Diversity
•Each symbol is transmitted L times
•The interval between symbol repetitions should be at least the coherence time,
not effective over slow fading channel
•Only one antenna is needed
•Reduction in efficiency (effective data rate < real data rate)

25. A. Chandra, NIT Durgapur – How to combat fading?
25
J.U., 13
th
April, 2007
Diversity Types (Contd.)
•Polarization Diversity
•Send signals with horizontal/ vertical polarization
•Compact co-located antennas
•Unequal branch powers, Less diversity gain, For fixed radio links
•Space-Time-Frequency Diversity
•Angle Diversity
•Field Component Diversity
•Antenna Pattern Diversity
•Multipath Diversity
•RAKE receiver
•Space-Time/ Space-Frequency/ Space-Time-Frequency Diversity
S. Kozono et al., ‘Base Station Polarization Diversity
Reception for Mobile Radio’, IEEE Trans. on Veh. Tech.,
vol. VT-33, no. 4, Nov., 1984.

26. A. Chandra, NIT Durgapur – How to combat fading?
26
J.U., 13
th
April, 2007
Control
Unit
Selection
Out
Selection Combining Equal Gain Combining
Out
Weights
and Phase
estimation
Weights
Phase
Maximal Ratio Combining
Out
Phase
estimation
Phase
Diversity Combining
î
í
ì
>
>
=
122
211
;
;
YYY
YYY
Y
C
2
21YY
Y
C
+
=
2211 YaYaY
C +=
A sub-optimal version of selection combining is switch-and-stay combining in
which alternate antenna are chosen if signal falls below a certain threshold
Choose the best Simple average Weighted average

27. A. Chandra, NIT Durgapur – How to combat fading?
27
J.U., 13
th
April, 2007
Diversity: Some Facts
•More is Less!
As number of diversity paths (L) increases
we have diminishing improvement
•Think Optimum
Performance of combiners
SWC < SC < EGC < MRC
•If it’s Worse, it’s Better!
More improvement for Rayleigh fading
channel than Rician
•Correlation
If correlation ρ is non-zero still we have
sufficient improvement up to ρ < 0.5

28. A. Chandra, NIT Durgapur – How to combat fading?
28
J.U., 13
th
April, 2007
Diversity: Some Facts
•More is Less!
As number of diversity paths (L) increases
we have diminishing improvement
•Think Optimum
Performance of combiners
SWC < SC < EGC < MRC
•If it’s Worse, it’s Better!
More improvement for Rayleigh fading
channel than Rician
•Correlation
If correlation ρ is non-zero still we have
sufficient improvement up to ρ < 0.5
P. H. Phuong, ‘Analysis of Antenna Diversity Techniques
Used in MIMO System’, In Proc. of International
Symposium on Electrical & Electronics Engineering, Oct.,
2005, HCM City, Vietnam, pp.18-22.
BER for BPSK system with diversity order L=4

29. A. Chandra, NIT Durgapur – How to combat fading?
29
J.U., 13
th
April, 2007
Diversity: Some Facts
•More is Less!
As number of diversity paths (L) increases
we have diminishing improvement
•Think Optimum
Performance of combiners
SWC < SC < EGC < MRC
•If it’s Worse, it’s Better!
More improvement for Rayleigh fading
channel than Rician
•Correlation
If correlation ρ is non-zero still we have
sufficient improvement up to ρ < 0.5
P
e
vs. SNR for selected values of ρ for 8PSK with L = 4
E. Perahia & G. J. Pottie, ‘On Diversity Combining for
Correlated Slowly Flat- Fading Rayleigh Channels’, In
Proc. of IEEE International Conference on Serving
Humanity Through Communications, SUPERCOMM/ICC,
May 1994, pp.342-346

30. A. Chandra, NIT Durgapur – How to combat fading?
30
J.U., 13
th
April, 2007
Error Correction Coding
•For a given E
b
/N
0
, with coding present, the error floor out of the demodulator
will not be lowered, but a lower error rate out of the decoder can be achieved
•For a given error performance, a code reduces the required E
b
/N
0
•Effective data rate decreases
•Coding Types

Block Code  Hadamard, Golay, BCH, RS
Convolutional Code  Viterbi Decoding
Turbo Code (Berrou ‘93)  Shannon limit
TCM (Ungerboeck ‘87)  Joint coding & modulation
E
b
/N
0
necessary for P
e
=10
-5
as a
function of code rate R
J. Hagenauer et al., ’Forward Error
correction Coding for Fading
Compensation in Mobile Satellite
Channels’, IEEE JSAC, vol. 5, no. 2,
Feb 1987, pp. 215-225

31. A. Chandra, NIT Durgapur – How to combat fading?
31
J.U., 13
th
April, 2007
Fading in Wireless Channels
Mitigation of Slow Flat Fading
Mitigation of Fast Fading
Equalization
OFDM
Modulation
FH/SS
RAKE Receiver
Mitigation of Frequency Selective Fading

32. A. Chandra, NIT Durgapur – How to combat fading?
32
J.U., 13
th
April, 2007
Frequency Selective Fading Mitigation
•Equalization
•Frequency selective channel introduce different attenuation & phase shift
to different frequency components in transmitted signal
•Equalizer does the opposite
•Frequency selective channel appears as flat fading channel
•Decision Feedback Equalizer (DFE)
•Maximum Likelihood Sequence Estimation
(MLSE) Equalizer
When a symbol is detected, the ISI it introduce on future symbols are
estimated and subtracted before the detection of subsequent symbols
•All possible data sequences are tested  optimum case
•Viterbi Equalizer applied to GSM

33. A. Chandra, NIT Durgapur – How to combat fading?
33
J.U., 13
th
April, 2007
•Orthogonal Frequency Division Multiplexing (OFDM)
Freq. Select. Fading Mitigation (Contd.)
•Available bandwidth is divided into several narrow band carriers
•Serial data stream is divided in N parallel data streams
•Fast serial data stream is transformed into slow parallel data streams 
Longer symbol durations  A frequency selective channel appears as flat
•Cyclic Prefix is inserted between consecutive OFDM symbols  removes
ISI from previous symbol
•Much more sensitive to synchronization errors, High peak to average
power ratio, Wastage of bandwidth in cyclic prefix
1 2 3 N-1 N
f
W
W/N
f
CP
CP
ISI

34. A. Chandra, NIT Durgapur – How to combat fading?
34
J.U., 13
th
April, 2007
J. Chuang, ’The Effects of Time Delay Spread on Portable
Radio Communications Channels with Digital Modulation’,
IEEE JSAC, vol. 5,no. 5, Jun’87, pp. 879-889
The irreducible BER performance for different
modulations plotted against rms delay spread
normalized by bit period.
•Choice of Modulation
•Pilot signal assisted Modulation
Freq. Select. Fading Mitigation (Contd.)
•4-ary modulations (QPSK, OQPSK, MSK) are
more resistant to delay spread than BPSK for
constant information throughput
•4-ary keying is used widely in 2G & 3G
•Frequency Hopping
Spread Spectrum (FH/SS)
•Receiver frequency band is changed
before the arrival of the multiple diffused
components
•Facilitate coherent detection
•Freq domain  In-band tones
•Time domain  Digital sequences

35. A. Chandra, NIT Durgapur – How to combat fading?
35
J.U., 13
th
April, 2007
Freq. Select. Fading Mitigation (Contd.)
•RAKE Receiver
A receiver technique which uses several baseband
correlators to individually process several signal multipath
components. The correlator outputs are combined to
achieve improved communications reliability and
performance.
RAKE
?
R. Price & P.E. Green, ’A Communication Technique for Multipath Channels’,
Proc. IRE, vol. 46, 1958, pp. 555-570
•IS-95
•Base station combines outputs
of its RAKE-receiver fingers (4 to
5) non-coherently
•Mobile receiver combines its
RAKE-receiver finger (generally
3) outputs coherently

36. A. Chandra, NIT Durgapur – How to combat fading?
36
J.U., 13
th
April, 2007
Fading in Wireless Channels
Mitigation of Slow Flat Fading
Mitigation of Frequency Selective Fading
Coding & Interleaving
Signal Redundancy
Robust Modulation
Doppler Diversity
Mitigation of Fast Fading

37. A. Chandra, NIT Durgapur – How to combat fading?
37
J.U., 13
th
April, 2007
Fast Fading Mitigation
X X X X X X
A
1
A
2
A
3
A
4
A
5
A
6
B
1
B
2
B
3
B
4
B
5
B
6C
1
C
2
C
3
C
4
C
5
C
6
D
1
D
2
D
3
D
4
D
5
D
6
E
1
E
2
E
3
E
4
E
5
E
6
F
1
F
2
F
3
F
4
F
5
F
6
A B C D E F
A
1
A
2
A
3
A
4
A
5
A
6
B
1
B
2
B
3
B
4
B
5
B
6C
1
C
2
C
3
C
4
C
5
C
6
D
1
D
2
D
3
D
4
D
5
D
6
E
1
E
2
E
3
E
4
E
5
E
6
F
1
F
2
F
3
F
4
F
5
F
6
A B C D E F
A
1
B
1
C
1
D
1
E
1
F
1
A
2
B
2
C
2
D
2
E
2
F
2
A
6
B
6
C
6
D
6
E
6
F
6
A
3
B
3
C
3
D
3
E
3
F
3A
4
B
4
C
4
D
4
E
4
F
4
A
5
B
5
C
5
D
5
E
5
F
5
1 2 3 4 5 6
Error burst
•Original code words
•De-interleaved words
•Interleaved words
•Coding & Interleaving
•Wireless multipath channels have memory  multiple copies of a symbol
arrive in delayed fashion and affect future symbol
•Even with fast fading several successive symbol transmissions are affected
 Burst Errors
•FEC schemes are designed for isolated errors, not for Burst Errors.

38. A. Chandra, NIT Durgapur – How to combat fading?
38
J.U., 13
th
April, 2007
R. Padovani, Reverse Link Performance of 1S-95 Based
Cellular Systems’, IEEE Personal Communications, vol. 1,
no. 3, 3rd qtr. 1994, pp. 28 - 34
Typical E
b
/N
0
,performance vs. vehicle speed for 850
MHz links to achieve a FER = 1% over a Rayleigh
channel with two independent paths
•Coding & Interleaving (Contd.)
•Interleaving separate symbols in an error
burst and spread them over time
•If the time separation is more than coherence
time, errors are uncorrelated in time
•Channel can be viewed as memoryless
•Interleaving realizes time diversity
Fast Fading Mitigation (Contd.)
As the motion increases in velocity, so does
the benefit of a given interleaver to the error
performance of any system

39. A. Chandra, NIT Durgapur – How to combat fading?
39
J.U., 13
th
April, 2007
•Signal Redundancy
•Fast fading occurs in low data rate transmission
•If symbol duration is reduced compared to coherence time, the channel
appears as slow fading channel
•Robust Modulation
•Non-coherent or, differentially coherent modulation
•Phase tracking not required  Detector integration time reduces
•Doppler Diversity
•Doppler spread induced by temporal channel variations can provide another
means for diversity that can be exploited to combat fading
•Applicable to CDMA spread-spectrum RAKE receiver
Fast Fading Mitigation (Contd.)
A. M. Sayeed & B. Aazhang, ‘Joint Multipath-Doppler Diversity in Mobile Wireless
Communications’, IEEE Trans. on Commun., vol. 47, no. 1, 1999, pp 123-132.

40. A. Chandra, NIT Durgapur – How to combat fading?
40
J.U., 13
th
April, 2007
•New/ Hybrid Technologies?
•Space Time Coding
•BLAST
•UWB
•MIMO-OFDM
•Cognitive Radio – Radio with Brain?
Cognitive radios will have the ability of devices to determine their location,
sense spectrum use by neighboring devices, change frequency, adjust
output power, and even alter transmission parameters and characteristics.
Future???