TE 4103 SISTEM KOMUNIKASI BERGERAK

Modul 6 Teknik-Teknik Mengatasi Fading

Jurusan Teknik Elektro Program Studi

S1
INSTITUT TEKNOLOGI TELKOM
2008
Typical Mobile Radio Propagation Channel

Amplitude

Frequency
Time
 Fading Mechanisms
Time dispersion
Time variations of the channel are caused by motion of the antenna
Channel changes every half a wavelength
Moving antenna gives Doppler spread
Fast fading requires short packet durations, thus high bit rates
Time dispersion poses requirements on synchronization and rate of convergence
of channel estimation
Interleaving may help to avoid burst errors

Frequency dispersion
Delayed reflections cause intersymbol interference (ISI)
Channel Equalization may be needed.
Frequency selective fading
Multipath delay spreads require long symbol times
Frequency diversity or spread spectrum may help

RSL Fluctuation
Shadowing, obstruction, etc
 Time Dispersion and Frequency Dispersion

Time Dispersion
Frequency Dispersion

Time Domain Channel variations Delay spread

Interpretation Fast Fading InterSymbol Interference
Correlation Distance Channel equalization

Frequency Doppler spectrum Frequency selective fading

Domain Intercarrier Interference Coherence bandwidth
Interpretation
 Effect of Fading
Freq. Selective Fading
Spectral density

TX BW > Channel BW
Bs Bs > Bc

Bc Freq.
Freq. Flat Fading
Spectral density

TX BW < Channel BW
Bs < B c
Bs

Coherent BW, Bc Freq.

 Statistical Fluctuations
Area-mean power
is determined by path loss
is an average over 100 m - 5 km
Local-mean power
is caused by local 'shadowing'
effects
has slow variations
is an average over 40 (few
meters)
Instantaneous power
fluctuations are caused by multipath reception
depends on location and frequency
depends on time if antenna is in motion
has fast variations (fades occur about every half a wave length)
 Fading Mitigation Techniques
Three techniques commonly use to
combat the effect of fading without
increasing TX Power and BW: diversity,
channel encoding, and equalization.
While Fading margin and Power control
are used to maintain a good signal
reception at Receiver.
 Fading Mitigation Technique:
( Diversity )
Diversity exploits the random nature of radio
propagation by finding the independent signal
paths. If one path undergo a deep fade, another
path may have a strong signal.
Usually employed to reduce the depth and
duration of fade experienced by receiver in flat
fading channel.
Types of diversity: spatial, frequency, time, and
polarization
 Spatial Diversity
Use two or more receiving antenna
While one antenna sees a null signal, the
others may receive a peak signals.
The received signals are then combined and
processed by an algorithm to get best
reception.
Can be implemented in both BS and MS
receiver
 Spatial Diversity

ro(t) Antenna is spaced

wo each other by an odd
d
r1(t) y(t) integer multiply of
w /4, usually d > 8 .
1 Spatial diversity can
rK(t) wK improve SNR at
receiver by as much
Processor
as 20 dB to 30 dB.

Combining algorithm commonly used: Selective, Equal gain,

and Maximal ratio combining.
 Selective Combiner
Ant. 1

G1

Switching Logic
Ant. 2 output
G2 or
Demodulator

Ant. m
Gm

Variable gain
 Selective Combining
Receiver only select one strongest signal to detect.
If average SNR of received signal in a branch = and
threshold SNR = , then probability that M branches of
antenna receive signals with SNR below the threshold is:

P(i < ) = PM(1 - e-

In other word, probability that received signal SNR above

the threshold is :

P(i > ) = 1 - PM(1 - e-

 Selective Combining
Example: 4 antenna diversity is used. If average SNR is 20 dB,
determine the probability that SNR will drop below 10 dB (bad
reception), and also that good reception (SNR above 10 dB) will mostly
take place. Compare with single antenna receiver!
Answer:
Threshold SNR = = 10 dB, = 20 dB, = 0.1

P4(i< 10 dB) = (1 e-0.1)4 = 0.000082, and

P4(i> 10 dB) = 1- (1 e-0.1)4 = 0.999918 or 99.9918%
With single antenna:

P (i< 10 dB) = (1 e-0.1) = 0.095, and

P (i> 10 dB) = 1- (1 e-0.1) = 0.905 or 90.5%

Improvement factor about 3 order in magnitude!

Max. Ratio Combiner

Ant. 1

G1

Co-phase
Ant. 2 M output
G2 and Detector
Sum
Ant. m m
Gm

Adaptive control
Variable gain
 Max. Ratio Combining
Signals from each branch/antenna are co-phased and
individually weighted to provide coherent addition to get
optimal SNR.
Probability that received signal SNR below threshold is:
(M
/ ) k 1
1 P M 1 e /
k 1 ( k 1)!

Probability of good reception:

M
( / ) k 1
P M 1 e /
k 1 ( k 1)!
 Equal Gain Combining
If weight of each branch is set to unity and
co-phased, Max. ratio combining become
equal gain combining.
Less complex with slightly lower
performance than max. ratio combining.
Without continuously adapt each weight of
branches differently, it allows receiver to
exploit received signals simultaneously.
 CDMA RAKE Receiver
Z1
Correlator 1

r(t) Z2 Z Z m(t)
Correlator 2 2 Int. DC
IF or base band . .
CDMA signal . .
with multipath
Zm
components Correlator m m

Since chip rate of CDMA much greater than coherence BW, delay spread
merely provide a multiple delayed version of signals at receiver. Instead of
causing ISI, RAKE receiver attempts to collect multipath signals, process it by
separate correlator receiver, and combine the signals to have a better detection.
 C(t)

delay adj.
C(t-2)

korelator

BTS C(t-n)
 Frequency Diversity
Use two or more carrier frequency for transmission with spacing about
2 5 % f o.
Need to employ two or more Transmitter and Receiver
Improvement factor :
 Channel Equalizer
Channel equalizer is employed to compensate ISI.
Since multipath fading channel is dynamic, equalizer must
be adaptive
(i) (i-D) (i-D-v) (i-D-V+1)
i time index
z-1
z-1
z-1
V equalizer order
D delay index
b0 bD+v bV-1

Adaptive
algorithm

(i)out
(i)
 Types of Equalizer
Linear:
Transversal filter (Zero forcing, LMS, RLS,
fast RLS, Sq. root RLS)
Lattice Filter (Gradient RLS)
Non Linear:
DFE (LMS, RLS, Fast RLS, Sq. root RLS)
ML Symbol Detection
MLSE
 Time Diversity
Interleaver
Read out bits to modulator one row at a time

1 m+1
2 m+2
Read in m
Coded bits
from . r
encoder o
w
s
M 2m nm

n columns
 Channel Encoding
Channel encoding is done by encode the data into a
special form, and introduce redundancies in the
transmitted data.
It protects data/information from error and distortion
introduced by the channel.
Redundant bits increase data rate hence the bandwidth,
but improve BER performance especially in fading
channel.
Reduce BW efficiency of the link in high SNR condition,
but provide excellent performance in low SNR condition
Two types mostly used: Block Code and Convolutional
code
 Fading Margin
Fading margin depends upon target availability of the link/
coverage.
Greater availability requires larger fading margin.
K u a t s in y a l ( d B ) s e t e la h
d it a m b a h f a d in g m a r g in ( F M )

T h e s h o ld FM

t
 Fading Margin
If fading margin FM applied to the link, then probability
that RSL at receiver separated at distance R above the
threshold can be written as:

11 11 FM

PPThThRR PP((m Th)) pp((m
mTh erf
dm erf
m))dm
FM

Th
Th
2 2
2 2 mm 2 2

Fading margin improve signal reception hence the

link performance, in an expense of increasing
transmission power.
 Power Control
Mitigating the effect of shadowing and near-far problem
If user 1 at 3 km from BTS transmitting with 100
mWatt, how much power is needed by user 2 at
9 km away from BTS using Okumura Hatta
model in urban area to achieve the same power
at the BTS with 10 m high above ground level?

d2
d1 Pr2

Pt2
Pt1 User 2

Answer: Path loss slope Hatta-Urban is( 44.9

Pr1
6.55 log 10) =38.35.
Basestation User 1
W2 = (d2/d1)3.835 W1 = 38.3 dBm =6.76 Watt
 Power Control
Rayleigh fading
Channel is estimated at the
30
Received signal amplitude receiver, then Tx is instructed
Controlled transmit power to adjust Tx power according to
Controlled SIR (target = 10 dB)
20 the estimated channel (e.g.
SNR).
10
Problem:
Signal level (dB)

Control rate >> fading rate

0
Control step size single step
or variable step
-10 What is the benefit/drawbacks
of single or variable step size ?
-20

-30
0 50 100 150 200 250 300
Time slot (0.67 ms)
Menjadi sangat fital karena sharing resource yang sama, setiap user
berprilaku sebagai random noise terhadap yang lain.

Pr1
Pt
Pt Pr1

user 2
user 1
 Power kontrol sangat bagus

Subc-2
Subc-1

5 km 10 km

Subc-1

Subc-2
 Distribusi power
power di penerima

User 1 S/N = 1/10

User 2 S/N = 10

frekuensi

Kualitas voice user 2 >> user 1, fenomena ini disebut problem

near-far. Dari dua user tadi kalau ditambahkan user ketiga
akan menurunkan kualitas baik user 3 maupun dua user yang
sudah ada.
Mengatasi problem near-far sehingga Pr(i) di BTS (up-link)
akan sama. Jika syarat S/N tiap user = 1/10, maka kapasitas
akan menjadi 11 user (meningkat signifikan).

Pr1 Pr2
Pt
Pt
Pr3 PrM

user 2
user 1

user 3
user M
power

User 1, S/N = 1/10

User 2, S/N = 1/10

User 3, S/N = 1/10

User 4, S/N = 1/10

User 5, S/N = 1/10

User 6, S/N = 1/10

User 7, S/N = 1/10

User 8, S/N = 1/10
User 9, S/N = 1/10
User 10, S/N = 1/10
User 11, S/N = 1/10
frekuensi

Dengan power kontrol kapasitas dimaksimalkan

 MS mengirim Pt maksimum ; pasti akan bisa diterima oleh
BTS, namun interferensi akan meningkat
MS mengirim Pt minimum ; ada kemungkinan tidak diterima
oleh BTS tetapi interferensi sangat kecil
IS-95 memberi solusi dengan :

Open loop mekanisme initial transmit power tanpa bantuan BTS

Close loop mekanisme pengendalian transmit power dengan

bantuan BTS
 Untuk mengatasi log-normal shadowing
Dengan access probe
Pt MS diestimasi dengan Pr MS

Pt(1) = -Pr - 73 +NOM_PWR +INIT_PWR IS-95

Pt n

Pr

Access probe
 PATH LOSS
Signal Level (dBm)

LOGNORMAL FADING (fading lambat)

RAYLEIGH FADING (fading cepat)

Distance (km)
 MS
trans. power

Koreksi
access probe
kedua

Koreksi
access probe
pertama
initial

Waktu
random random
Untuk mengatasi fluktuasi receive power karena Rayleigh
fading (fading cepat).
BTS memeriksa kualitas up-link secara kontinyu.
Jika kualitas jelek, BTS akan mengirim command(PCB)via
down-link untuk menaikan Pt MS.
Jika kualitas terlalu bagus akan dilakukan hal sebaliknya.
Eb/N0 digunakan sebagai indikator kualitas link.
PCB = +1dB jika Pt MS harus dinaikan dan -1dB jika
sebaliknya
Karena digunakan untuk mengatasi fading cepat, maka
proses pengendalian harus berlangsung cepat sehingga PCB

disisipkan pada kanal traffik arah down-link

 Power control (PCB)
800 bps

0 atau 1 tanpa proteksi

VOCODER KONVOLUSI MUX SPREADING

9,6 kbps 19,2 kbps 19,2 kbps 1,2288 Mcps

Power control
bit position
 20ms
PCG0 PCG1 PCG2 PCG3 PCG7 PCG12 PCG13 PCG14 PCG15

1,25 ms

Probabilitas meletakan PCB

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

PCB = 0 MS harus naik 1 dB

PCB = 1 MS harus turun 1 dB
1 0 1 1

CONTROL POSITION
(11)10
 Eb/N0 diukur di BTS

UL
PCG0 PCG1 PCG2 PCG3 PCG4 PCG5 PCG6 PCG7 PCG8

Selama menerima PCG8

BTS memutuskan kirim 1 atau 0

DL
PCG0 PCG1 PCG2 PCG3 PCG4 PCG5 PCG6 PCG7 PCG8 PCG9

BTS MENGIRIM
PCB
 FTC TRANSMITTER

KE MS

PCB

DARI MS

ESTIMASI
Eb/N0

DECISION
RX EST >Th PCB = 1
EST < Th PCB = 0

ESTIMASI Th. Eb/N0

DEM
FER

BASE STATION
TRANSMITTER RTC

Pt yang harus dipakai

selama 1,25 ms
DECISION
PCB = 0 , +1dB
+1 dB atau -1 dB PCB = 1, -1dB
DEM RX

INIT_PWR (dB)

NOM_PWR (dB) TCH

K= -73 (dB)

ESTIMASI
(-1) RX TOTAL
-Pr Pr

MOBILE STATION
Some practical approaches
 How to handle fast multipath fading
Analog
User must speak slowly
GSM
Error correction and interleaving to avoid
burst errors
Error detection and speech decoding
Fade margins in cell planning
DECT
Diversity reception at base station
IS95 (CDMA2000)
Wideband transmission averages channel
behaviour
This avoids burst errors and deep fades
 How do systems handle Doppler spreads?
Analog
Carrier frequency is low enough to avoid problems
GSM
Channel bit rate well above Doppler spread
TDMA during each bit / burst transmission the channel is fairly
constant.
Receiver training/updating during each transmission burst
Feedback frequency correction
DECT
Only small Doppler spreads are to be anticipated for
Original DECT concept did not standardize an equalizer
IS95 (CDMA2000)
Downlink: Pilot signal for synchronization and channel estimation
Uplink: Continuous tracking of each signal
How do systems handle delay spreads?
Analog
Narrowband transmission
GSM
Adaptive channel equalization
Channel estimation training sequence
DECT
Use the handset only in small cells with small delay spreads
Diversity and channel selection can help a little bit
pick a channel where late reflections are in a fade
IS95
Rake receiver separately recovers signals over paths with excessive delays
Digital Audio Broacasting
OFDM multi-carrier modulation
The radio channel is split into many narrowband (ISI-free) subchannels