This action might not be possible to undo. Are you sure you want to continue?
DIVISION MULTIPLEXING
FOR WIRELESS CHANNELS
LEONARD J. CIMINI, JR.
YE (GEOFFREY) LI
AT&T LABS  RESEARCH
ORTHOGONAL FREQUENCY DIVISION
MULTIPLEXING FOR WIRELESS
CHANNELS
Leonard J. Cimini, Jr.
AT&T Labs – Research
100 Schulz Dr., Rm. 4146
Red Bank, NJ 077017033, USA
Email: ljc@research.att.com
Tel/Fax: 17323453129/3039
Ye (Geoffrey) Li
AT&T Labs – Research
100 Schulz Dr., Rm. 4152
Red Bank, NJ 077017033, USA
Email: ljye@research.att.com
Tel/Fax: 17323453132/3039
ABSTRACT
Orthogonal frequency division multiplexing (OFDM) has been shown
to be an effective technique to combat multipath fading in wireless
communications. It has been successfully used for HF radio
applications and has been chosen as the standard for digital audio
broadcasting and digital terrestrial TV broadcasting in Europe and
highspeed wireless local areas networks. In this tutorial, we present
the basic principles of OFDM and discuss the problems, and some of
the potential solutions, in implementing an OFDM system. Techniques
for peaktoaverage power ratio reduction, time and frequency
synchronization, and channel estimation will be discussed. We
conclude with a brief overview of current application areas
BIOGRAPHIES OF PRESENTERS:
Leonard J. Cimini, Jr., received the B.S.E. (summa cum laude), M.S.E.
and Ph.D. degrees in electrical engineering from the University of
Pennsylvania in 1978, 1979, and 1982, respectively. During the
graduate work he was supported by a National Science Foundation
Fellowship. Since 1982, he has been employed at AT&T, where his
research interests are in wireless communications systems. Dr. Cimini
is a member of Tau Beta Pi and Eta Kappa Nu. He has been very active
in the IEEE Communications Society and is EditorinChief of the IEEE
JSAC: Wireless Communications Series. He is also an Adjunct
Professor at the University of Pennsylvania.
Ye (Geoffrey) Li received the B.Eng and M.Eng degrees in 1983 and
1986, respectively, from the Department of Wireless Engineering,
Nanjing Institute of Technology, Nanjing, China, and the Ph.D. degree
in Electrical Engineering in 1994, Auburn University, Alabama. Since
May 1996, he has been with AT&T Labs  Research. His current
research interests are in statistical signal processing and wireless
communications. He has served as a guest editor for a special issue
on Signal Processing for Wireless Communications for the IEEE J
SAC and is an editor for Wireless Communication Theory for the IEEE
Transactions on Communications.
• Introduction
• Basic Concepts
• PeaktoAverage Power Ratio Reduction
• Time and Frequency Synchronization
• Channel Estimation
• Applications
• Summary
• References
OUTLINE
• Introduction
• Basic Concepts
• PeaktoAverage Power Ratio Reduction
• Time and Frequency Synchronization
• Channel Estimation
• Applications
• Summary
• References
OUTLINE
INTRODUCTION
• Motivation
• Radio Environment
• Brief History
MOTIVATION
• Highbitrate wireless applications
• Limitations caused by the radio environment
• OFDM can overcome these inherent bit rate
limitations
PATH LOSS MODEL
• Path Loss
• Shadow Fading
• Multipath
• Flat fading
• Doppler spread
• Delay spread
• Interference
where P
r
is the local mean received signal
PATH LOSS MODEL
• Different, often complicated, models are
used for different environments.
• A simple model for path loss, L, is
The path loss exponent α αα α = 2 in free space;
2 ≤ ≤≤ ≤ α αα α ≤ ≤≤ ≤ 4 in typical environments.
power, P
t
is the transmitted power, and d is
the transmitter receiver distance.
α αα α
= == = = == =
d
1
K
P
P
L
t
r
SHADOW FADING
• The received signal is shadowed by
obstructions such as hills and buildings.
• This results in variations in the local mean
received signal power,
• Implications
– nonuniformcoverage
– increases the required transmit power
( (( ( ) )) ) ( (( ( ) )) )
( (( ( ) )) ) . dB 10 4 , , 0 N ~ G where
G dB P dB P
S
2
S S
S r r
≤ ≤≤ ≤ σ σσ σ ≤ ≤≤ ≤ σ σσ σ
+ ++ + = == =
MULTIPATH
Constructive and destructive interference
of arriving rays
Received
Power
Delay Spread
t
dB With Respect
to RMS Value
0 0.5
0.5λ λλ λ
1.5
30
20
10
10
0
1
t, in seconds
0 10 30 20
x, in wavelength
( (( ( ) )) ) ( (( ( ) )) )
i
j
i
i
t t e a t h
i
− −− − δ δδ δ = == =
θ θθ θ
∑ ∑∑ ∑
FLAT FADING
• The delay spread is small compared to the
symbol period.
• The received signal envelope, r, follows a
Rayleigh or Rician distribution.
• Implications
– increases the required transmit power
– causes bursts of errors
shadow fading
Rayleigh fading
path loss
log (distance)
Received
Signal
Power
(dB)
( (( ( ) )) ) r log 20
S
G ) dB (
r
P dB
r
P + ++ + + ++ + = == =
DOPPLER SPREAD
• A measure of the spectral broadening caused
by the channel time variation.
• Implications
– signal amplitude and phase decorrelate
after a time period ~ 1/f
D
Example: 900 MHz, 60 mph, f
D
= 80 Hz
5 GHz, 5 mph, f
D
= 37 Hz
λ λλ λ
≤ ≤≤ ≤
v
f
D
τ ττ τ
large
T
τ ττ τ
small
T
0
1 1
T 2T
Channel Input
Channel Output
0 T 2T
0 T 2T
Tworay model
τ ττ τ = rms delay spread
2τ ττ τ
Delay
R
e
c
e
i
v
e
d
P
o
w
e
r
τ ττ τ
T
• small negligible intersymbol interference
• large significant intersymbol interference,
which causes an irreducible error floor
τ ττ τ
T
DELAY SPREAD
TIME DOMAIN INTERPRETATION
B
s
= signal bandwidth ≈ ≈≈ ≈ 1/T
H(f)
B
s
1
2τ ττ τ
f
• small flat fading
τ ττ τ
T
• large frequencyselective fading
τ ττ τ
T
DELAY SPREAD
FREQUENCY DOMAIN INTERPRETATION
The rms delay spread imposes a limit on the
maximum bit rate. For example, for QPSK
• ISI causes an irreducible error floor.
τ ττ τ Maximum Bit Rate
Mobile (rural) 25 µ µµ µsec 8 kbps
Mobile (city) 2.5 µ µµ µsec 80 kbps
Microcells 500 nsec 400 kbps
Large Building 100 nsec 2 Mbps
+
x
+
+
+
+
+
x
x
x
x
x
10
2
10
4
10
3
10
2
10
1
10
1
10
0
BPSK
QPSK
OQPSK
MSK
Modulation
Coherent Detection
I
r
r
e
d
u
c
i
b
l
e
P
b
τ ττ τ
T
=
rms delay spread
symbol period
BIT RATE LIMITATIONS
INTERFRENCE
• Frequencies are reused often to maximize spectral
efficiency.
• For interferencelimited systems, the noise floor is
dominated by cochannel interference.
• Implications
– high reuse efficiency requires interference
mitigation
BASE
STATION
R
D
α αα α
= == = ≈ ≈≈ ≈
+ ++ + R
D
6
1
I
S
N I
S
• Military HF radio (1950’s  1960’s)
– Kineplex
– Kathryn
• Wireline modem (Telebit, Gandalf)
• Cellular modem (Telebit)
• Digital audio and terrestrial TV broadcasting
(Europe)
• Asymmetric digital subscriber line (DMT)
• Wireless LANs
– IEEE802.11  National Information
Infrastructure
– HIPERLAN TYPE II
HISTORY
• Introduction
• Basic Concepts
• PeaktoAverage Power Ratio Reduction
• Time and Frequency Synchronization
• Channel Estimation
• Applications
• Summary
• References
OUTLINE
BASIC CONCEPTS
• Multicarrier
• Basic OFDM
• Impairments
• Alternative forms
• The transmission bandwidth is divided into many
narrow subchannels which are transmitted in
parallel.
• Ideally, each subchannel is narrow enough so
that the fading it experiences is flat ⇒ ⇒⇒ ⇒no ISI.
RF
Transmitter
R/N b/s
D(t)
f
0
f
1
f
N1
d
N1
(t)
QAM
filter
R/N b/s
QAM filter
R/N b/s
QAM
filter
(t) d
0
(t) d
1
Bandlimited
signals
f
0
f
1
f
2
f
0
f
N1
f
1
N1
Receiver
RF
filter
QAM
QAM
QAM
filter
f
1
f
0
filter
f
N1
MULTICARRIER
• Disadvantage:
 Requires filter bank at receiver
 Spectrally inefficient
Horizontal slide here
• A guard interval can virtually eliminate ISI
BASIC OFDM RECEIVER
• Subchannel separation
– choose f
n
= f
0
+ n∆ ∆∆ ∆f, with ∆ ∆∆ ∆f =
– integrate over NT, then d(m) = d(m)
1
NT
^
(or, interblock interference) ⇒ ⇒⇒ ⇒lower spectral
or power efficiency.
parallel
to
serial
converter
∫ ∫∫ ∫
∫ ∫∫ ∫
∫ ∫∫ ∫
QAM
f
0
f
1
f
N1
d(0)
d(1)
d(N1)
RF
[ [[ [ ] ]] ]
s
t ) f k f ( 2 j
1 N
0 k
p
T t 0 , e k a ) t ( x
c
≤ ≤≤ ≤ ≤ ≤≤ ≤
ℜ ℜℜ ℜ = == =
∆ ∆∆ ∆ + ++ + π ππ π
− −− −
= == =
∑ ∑∑ ∑
[ [[ [ ] ]] ]
s
ft k 2 j
1 N
0 k
b
T t 0 , e k a ) t ( x ≤ ≤≤ ≤ ≤ ≤≤ ≤ = == =
∆ ∆∆ ∆ π ππ π
− −− −
= == =
∑ ∑∑ ∑
PASSBAND VERSUS BASEBAND
• Passband
• Baseband
DFT IMPLEMENTATION
TRANSMITTER
• Transmitted signal can be obtained
using a Discrete Fourier Transform
• If sampled at a rate of T
s
/N,
• For orthogonality, ∆ ∆∆ ∆fT
s
= 1,
• Efficient FFT implementation
[ [[ [ ] ]] ]
s
ft k 2 j
1 N
0 k
b
T t 0 , e k a ) t ( x ≤ ≤≤ ≤ ≤ ≤≤ ≤ = == =
∆ ∆∆ ∆ π ππ π
− −− −
= == =
∑ ∑∑ ∑
[ [[ [ ] ]] ] [ [[ [ ] ]] ]
N / fT nk 2 j
1 N
0 k
S b b
s
e k a T
N
n
x n x
∆ ∆∆ ∆ π ππ π
− −− −
= == =
∑ ∑∑ ∑
= == =
= == =
[ [[ [ ] ]] ] [ [[ [ ] ]] ] [ [[ [ ] ]] ] { {{ { } }} } k a IDFT e k a n x
N / nk 2 j
1 N
0 k
b
= == = = == =
π ππ π
− −− −
= == =
∑ ∑∑ ∑
DFT IMPLEMENTATION
RECEIVER
[ [[ [ ] ]] ] [ [[ [ ] ]] ] { {{ { } }} }
[ [[ [ ] ]] ]
[ [[ [ ] ]] ]
[ [[ [ ] ]] ]
[ [[ [ ] ]] ] [ [[ [ ] ]] ]
[ [[ [ ] ]] ] k a
k m N m a
N
1
e m a
N
1
e m a
N
1
e n x
N
1
n x DFT k a
ˆ
1 N
0 m
1 N
0 n
N / ) k m ( n 2 j
1 N
0 m
N / ) k m ( n 2 j
1 N
0 n
1 N
0 m
N / nk 2 j
1 N
0 n
b
b
= == =
− −− − δ δδ δ = == =
= == =
= == =
= == =
= == =
∑ ∑∑ ∑
∑ ∑∑ ∑ ∑ ∑∑ ∑
∑ ∑∑ ∑∑ ∑∑ ∑
∑ ∑∑ ∑
− −− −
= == =
− −− −
= == =
− −− − π ππ π
− −− −
= == =
− −− − π ππ π
− −− −
= == =
− −− −
= == =
π ππ π − −− −
− −− −
= == =
• Coding across subchannels ⇒ ⇒⇒ ⇒works best with
large delay spread
• Adaptive loading
– More bits/symbol where SNR is sufficient
– Could also adapt transmit power in each
subchannel
– Requires reliable feedback channel and
accurate channel information
PERFORMANCE IMPROVEMENT
• Frequency equalization and coherent detection
⇒ ⇒⇒ ⇒requires accurate channel estimation
SAMPLE DESIGN
• Goal
– Transmit 1.2 Mbits/sec using QPSK with
B=800 kHz bandwidth channel
– Delay span up to 40 µ µµ µsec (max 5 kbaud for
single carrier)
• Design
– Choose subchannel width so that there is no
ISI in each subchannel ⇒ ⇒⇒ ⇒∆ ∆∆ ∆ f = 6.25 kHz ⇒ ⇒⇒ ⇒N =
B/∆ ∆∆ ∆f = 128 subchannels
– OFDM symbol duration T
s
=1/∆ ∆∆ ∆f = 160 µ µµ µsec
– Guard interval T
g
= 40 µ µµ µsec
– OFDM block length: T
f
= T
s
+ T
g
= 200 µ µµ µsec
– Assuming 4 guard channels on each end, there
are 120 data subchannels, each transmitting 2
bits in 200 µ µµ µsec
sec / Mbits 2 . 1
sec 200
bits 2 x 120
R
b
= == =
µ µµ µ
= == =
IMPAIRMENTS
• Timevarying fading, frequency offset, and
timing mismatch impair the orthogonality of
the subchannels.
• Large amplitude fluctuations can be a serious
problem when transmitting through a
nonlinearity.
TIMEVARYING IMPAIRMENTS
• General expression:
• Frequency offset
– For a frequency offset between the
transmitter and receiver,
[ [[ [ ] ]] ] [ [[ [ ] ]] ] { {{ { } }} } [ [[ [ ] ]] ] [ [[ [ ] ]] ] [ [[ [ ] ]] ] [ [[ [ ] ]] ]
O O N O OM L
ON OM L
ICI
rotated & attenuated
k , n K n a k , k K k a n DFT k a
ˆ
k n
b ∑ ∑∑ ∑
≠ ≠≠ ≠
+ ++ + = == = χ χχ χ = == =
( (( ( ) )) ) [ [[ [ ] ]] ]
( (( ( ) )) )
Ts t 0 , e k a t x
t f f k 2 j
1 N
0 k
b
≤ ≤≤ ≤ ≤ ≤≤ ≤ = == =
δ δδ δ − −− − ∆ ∆∆ ∆ π ππ π
− −− −
= == =
∑ ∑∑ ∑
[ [[ [ ] ]] ]
[ [[ [ ] ]] ]
f
f
j
e
f
f
f
f
sin
k , k K
f
f
k n j
e
f
f
k n
f
f
k n sin
k n, K
∆ ∆∆ ∆
δ δδ δ
π ππ π
] ]] ]
] ]] ]
] ]] ]
] ]] ]
] ]] ]
] ]] ]
∆ ∆∆ ∆
δ δδ δ
π ππ π
  
¹ ¹¹ ¹
  
' '' '
\ \\ \
  
∆ ∆∆ ∆
δ δδ δ
π ππ π
= == =
∆ ∆∆ ∆
δ δδ δ
− −− − − −− − π ππ π
] ]] ]
] ]] ]
] ]] ]
] ]] ]
] ]] ]
] ]] ]
  
¹ ¹¹ ¹
  
' '' '
\ \\ \
  
∆ ∆∆ ∆
δ δδ δ
− −− − − −− − π ππ π
] ]] ]
] ]] ]
] ]] ]
  
¹ ¹¹ ¹
  
' '' '
\ \\ \
  
∆ ∆∆ ∆
δ δδ δ
− −− − − −− − π ππ π
= == =
TIMING MISMATCH
• Timing offset smaller than the guard
interval results in a phase shift.
⇓ ⇓⇓ ⇓
a phase shift
• Otherwise, additional interference is
generated.
• Best solution is to choose sufficient
guard interval.
[ [[ [ ] ]] ] [ [[ [ ] ]] ]
t f 2 j
t
k
e k a k a
ˆ
δ δδ δ π ππ π
δ δδ δ
= == =
DELAY SPREAD
• Assuming timeinvariance, the
multipath channel results in a
received signal at kth subchannel
• Simple complex multiplicative
distortion
• For coherent detection, channel
parameter estimation and tracking
are required.
[ [[ [ ] ]] ] [ [[ [ ] ]] ] [ [[ [ ] ]] ] [ [[ [ ] ]] ] k w k a k H k a
ˆ
+ ++ + = == =
NONLINEARITIES
• Large peaktoaverage power ratio (PAPR)
• PAPR ~ number of subcarriers
• Large PAPR ⇒ ⇒⇒ ⇒inband distortion and spectral
spreading
• PAPR reduction techniques required
ALTERNATIVE FORMS
• Bandlimited OFDM:
– R.W. Chang (BSTJ, Dec. 1966)
– B.R. Saltzberg (IEEE Trans. on Comm.
Tech., Dec. 1967)
– B. Hirosaki (IEEE Trans. on Comm., Jan.
1980)
• WaveletBased OFDM:
– B. LeFloch (Proc. IEEE, June 1995)
• Introduction
• Basic Concepts
• PeaktoAverage Power Ratio Reduction
• Time and Frequency Synchronization
• Channel Estimation
• Applications
• Summary
• References
OUTLINE
PAP RATIO REDUCTION
• PAPR properties
• Clipping and filtering
• Selective mapping
• Partial transmit sequences
• Coding
• Other techniques
• Superposition of a large number of subcarrier
signals results in a Rayleigh envelope.
• PAPR definition
For N=128, PAPR = 21 dB.
• However, these large peaks do not occur very
often.
PAPR PROPERTIES
( (( ( ) )) )
N
P
t x max
PAPR
ave
2
b Ts t 0
= == = = == =
< << < < << <
PAPR PROPERTIES
• Large PAPR
– In band noise ⇒ ⇒⇒ ⇒increases BER
– Spectral spreading ⇒ ⇒⇒ ⇒ACI
• Possible Solutions
– Amplifier backoff
– Reduce PAPR of OFDM signal
• Deliberate clipping will reduce peak value,
but will result in spectral spreading (ACI)
and inband distortion (BER).
• Filtering is required to minimize spectral
spreading. ⇒ ⇒⇒ ⇒peak regrowth
CLIPPING AND FILTERING
CLIPPING AND FILTERING
10
0
10
1
10
2
10
3
10
4
0 2 4 6 8 10 12 14 16
SNR (dB)
Unclipped
CR=0.8
1.0
1.2
1.4
1.6
B
E
R
0
P
S
D
(
d
B
)
CR = 1.4
10
20
30
40
50
60
70
80
90
10
4 3 2 1 0 1 2 3
NORMALIZED FREQUENCY
4
• Multiply data signal by M different sequences,
r
1
,… r
M
,
• Convert each data sequence into the time
domain with an Npoint IFFT
• Select sequence for transmission with the
smallest PAPR
SELECTIVE MAPPING
• Divide the OFDM tones into M clusters
• Convert each cluster into the timedomain
using an Npoint IFFT
• Combine the M output sequences to
minimize the PAPR
PARTIAL TRANSMIT SEQUENCE
Using the same redundancy, PTS can achieve a
lower PAPR at the expense of more complexity.
PERFORMANCE
• Nonlinear
– Map transmitted sequence into a larger
sequence where highpeak sequences
are not used
– Good performance with little overhead
– Requires table lookup ⇒ ⇒⇒ ⇒only applicable
for small number of subchannels
– Error propagation
• Current work searching for systematic
implementation with some error correction
capability
CODING
• Introduction
• Basic Concepts
• PeaktoAverage Power Ratio Reduction
• Time and Frequency Synchronization
• Channel Estimation
• Applications
• Summary
• References
OUTLINE
TIME AND FREQUENCY
SYNCHRONIZATION
• Timing offset estimation
• Frequency offset estimation
• Joint offset estimation
TIMING OFFSET ESTIMATION
• Pilotbased methods
– NonOFDMbased pilot symbols
– OFDMbased pilot symbols
• Nonpilot based methods
PILOTBASED TIMING
NonOFDM Pilot Symbols
Use a null signal inserted at the start of each
group of OFDM blocks
OFDM OFDM OFDM OFDM OFDM OFDM
Null
Symbol
Null
Symbol
time
. . .
PILOTBASED TIMING
OFDM Pilot Symbols
Design special OFDM block for estimation (Moose,
Schmidl)
. . .
Cyclic
prefix
= time
frequency
N tones
First Half of Symbol Second Half of Symbol
PILOTBASED TIMING
Performance
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
2 4 6 8 10 12 14 16 18 20
SNR (dB)
p
r
o
b
a
b
i
l
i
t
y
t
i
m
e

o
f
f
s
e
t
e
s
t
i
m
a
t
o
r
o
u
t
s
i
d
e
s
a
l
e

r
e
g
i
o
n
o
f
p
i
l
o
t

r
e
g
i
o
n
'ML' pilot timeoffset estimator
Schmidl estimator
NONPILOT BASED TIMING
Use redundancy in the cyclic prefix to
estimate the time offset
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.1
2 4 6 8 10 12 14 16 18 20
SNR (dB)
p
r
o
b
a
b
i
l
i
t
y
t
i
m
e

o
f
f
s
e
t
e
s
t
i
m
a
t
o
r
o
u
t
s
i
d
e
s
a
l
e
r
e
g
i
o
n
ML
Squareddifference
Correlation only
FREQUENCY OFFSET ESTIMATION
• Pilotbased methods
– NonOFDMbased pilot symbols
– OFDMbased pilot symbols
• Nonpilot based methods
OFDM symbol
with pilots
OFDM symbol OFDM symbol
OFDM symbol
with pilots
D1 OFDM symbols
time
Frequency domain of an OFDM pilot symbol
...
...
Pilot tones spaced every 4 tones.
Pilot tone
FREQUENCY OFFSET ESTIMATION
PilotBased
An OFDMbased pilot scheme for coarse and fine
frequency synchronization
FREQUENCY OFFSET ESTIMATION
Performance
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
2 4 6 8 10 12 14 16 18 20
SNR (dB)
p
r
o
b
a
b
i
l
i
t
y
o
f
i
n
c
o
r
r
e
c
t
t
o
n
e

o
r
d
e
r
i
n
g
epsilon = 0
epsilon = .1
epsilon = .2
epsilon = .3
JOINT ESTIMATION
Based on shortened Moose pilot symbol
1
2
3
4
5
6
7
8
2 4 6 8 10 12 14 16 18 20
SNR (dB)
s
t
a
n
d
a
r
d
d
e
v
i
a
t
i
o
n
o
f
t
i
m
e

o
f
f
s
e
t
e
s
t
i
m
a
t
o
r
'ML' pilotbased estimator
Schmidl estimator
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
2 4 6 8 10 12 14 16 18 20
SNR (dB)
'ML' pilotbased estimator
Schmidl estimator
Estimated variance
JOINT ESTIMATION
Based on a 16sample cyclic prefix
0
5
10
15
20
25
30
2 4 6 8 10 12 14 16 18 20
SNR (dB)
Joint ML
Squareddifference
Shortened squared difference
0
0.01
0.02
0.03
0.04
0.05
0.06
2 4 6 8 10 12 14 16 18 20
SNR (dB)
Joint ML
Squareddifference
Shortened squared difference
Estimated variance
• Introduction
• Basic Concepts
• PeaktoAverage Power Ratio Reduction
• Time and Frequency Synchronization
• Channel Estimation
• Applications
• Summary
• References
OUTLINE
CHANNEL ESTIMATION
• Differential and coherent detection
• Decisiondirected estimation
• Pilotsymbolaided estimation
DIFFERENTIAL AND COHERENT
DETECTION
• 3dB SNR degradation for differential detection
• Differential
– in time ⇒ ⇒⇒ ⇒sensitive to Doppler shift
– in frequency ⇒ ⇒⇒ ⇒sensitive to delay spread
• Coherent detection requires channel
information
DECISIONDIRECTED ESTIMATION
• Use sliced data for estimating channel
parameters
• Obtain an MMSE or robust estimator using the
correlations of the channel parameters in time
and/or frequency.
DECISIONDIRECTED ESTIMATION
MMSE Estimator
• Using time and frequency correlations
• Estimator coefficients
• Estimator structure
H[n, 1]
H[n, 2]
H[n, 1]
H[n, 2]
H[n, K] H[n, K]
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
U
H
U
Φ ΦΦ Φ
1
(ω ωω ω)
Φ ΦΦ Φk
0
(ω ωω ω)
∼ ∼∼ ∼
∼ ∼∼ ∼
∼ ∼∼ ∼
^
^
^
[ [[ [ ] ]] ] [ [[ [ ] ]] ]
[ [[ [ ] ]] ] [ [[ [ ] ]] ] [ [[ [ ] ]] ] { {{ { } }} } n , ... , n diag n
U n U n C
N 1
H
Φ ΦΦ Φ Φ ΦΦ Φ = == = Φ ΦΦ Φ
Φ ΦΦ Φ = == =
DECISIONDIRECTED ESTIMATION
Robust Estimator
• Why robust estimation?
– A large performance degradation is
possible if MMSE estimator is not matched
to the channel.
• Robust design
– Match rectangular spectrum in time and
frequency domains
– Good performance for almost all channels
– Relatively insensitive to Doppler and delay
profiles
PILOTSYMBOLAIDED ESTIMATION
• Pilot symbol grid
• Obtain an estimate of the channel at the pilot
symbol positions.
• Obtain estimates at other frequencies and
times by interpolation.
PILOTSYMBOLAIDED ESTIMATION
Grid Selection
• 2D spectrum of channel parameters at pilot
symbol positions
• Nonrectangular pilot symbol grids are better
SIMULATION PARAMETERS
• 800 kHz bandwidth
• Number of subchannels N = 128
subchannels (4 guard subchannels at
each end)
• OFDM block duration T
f
= 200 µsec (with
40µsec guard interva)
• (40,20) RS code, which corrects 10
erasures, based on signal strength, and
correct 5 random errors
WORD ERROR RATE
Typical Urban Channel
WORD ERROR RATE
Hilly Terrain Channel
• Introduction
• Basic Concepts
• PeaktoAverage Power Ratio Reduction
• Time and Frequency Synchronization
• Channel Estimation
• Applications
• Summary
• References
OUTLINE
APPLICATIONS
• Asymmetric digital subscriber line
• Digital audio and terrestrial TV
broadcasting
• Wireless LAN’s
• Highspeed cellular data
DIGITAL AUDIO BROADCASTING
• Broadcasting standard in Europe
• Single frequency network
• Bandwidth = 7 MHz
• Useful bit rate = 5.6 Mbits/sec
• N = 448 subchannels
• T
f
= 80 µ µµ µsec, T
f
= 64 µ µµ µsec, and T
g
= 16 µ µµ µsec
• Rate1/4 conv. coding with constraint length 7
• Time interleaving: 16 X 24 msec
• Pilot symbols for channel estimation
WIRELESS LANS
• IEEE802.11(a)
– New NII spectrum at about 5 GHz
– Indoor applications
OFDM
Convolution K=7, R=1/2 or R=3/4
intercarrier interleave
Modulation scheme
DBPSK, 16 QAM in each subchannel Subcarrier Modulation
48 Subchannels out of 64 Number of subchannels
5 Mbps (BPSK, R=1/2) Data rates
Coding
10 Mbps (QPSK, R=1/2)
15 Mbps (QPSK, R=3/4))
20 Mbps (16QAM, R=1/2)
30 Mbps (16QAM, R=3/4)
WIRELESS LANS
48, 32, 24, 16 and 8 Mbps Information data rate
• HIPERLAN: High Performance Radio Local
Area Network
– Standard in Europe at 5.2 GHz
– HIPERLAN Type I: single carrier
GSMK with equalization
– HIPERLAN Type II: COFDM
OFDM with 16QAM, QPSK or BPSK Modulation
Convolutional 3/4 or 1/2 Coding rate
48 Coding
3 µ µµ µs OFDM symbol duration
600 ns Guard interval
600 ns T
prefix
75 ns T
postfix
416.666 kHz Subchannel spacing
0.025 Rolloff factor
25 MHz Channel Spacing
20 MHz Occupied 3 dB Bandwidth
Band Division Multiple Access (BDMA)
• Proposal by Sony, Japan
• Combination of TDMA, OFDM and cluster
hopping
HIGH SPEED CELLULAR
• Parameters
– number of tones per cluster: N
c
= 24
– tone spacing: ∆ ∆∆ ∆f = 4.17 kHz
– symbol duration: T
s
= 1/∆ ∆∆ ∆f = 240 µsec
– guard interval: T
g
= 38.8 µsec
– ramp time: T
r
= 10 µsec
– block length: T
f
= T
s
+ T
r
+ T
g
= 288.5 µsec
– cluster width: B
c
= N
c
∆ ∆∆ ∆f = 100 kHz
HIGH SPEED CELLULAR
• Goal to provide widearea Internet service
to mobile subscribers
• Combines OFDM with multiple transmitter
and receiver antennas and coding
• Parameters
– total bandwidth: B = 800 kHz
– number of tones: N = 192
– tone spacing: ∆ ∆∆ ∆f = B/N = 4.17 kHz
– symbol duration: T
s
= 1/∆ ∆∆ ∆f = 240 µ µµ µsec
– guard interval: T
g
= 48.5 µ µµ µsec
– block length: T
f
= T
s
+ T
g
= 288.5 µ µµ µsec
– rate1/2 RS code across subchannels
Advanced Cellular Internet Service (ACIS)
• Introduction
• Basic Concepts
• PeaktoAverage Power Ratio Reduction
• Time and Frequency Synchronization
• Channel Estimation
• Applications
• Summary
• References
OUTLINE
• Highbitrate wireless data is desirable, but
the radio environment puts an upper limit
on the achievable bit rate.
• OFDM, by transmitting data over many
narrow subchannels, can overcome the bit
rate limit.
• However, to realize an OFDM system,
several practical issues must be addressed,
including PAPR, frequency offset and
timing mismatch, and channel estimation.
• Several promising solutions have been
proposed for all of these problems.
• OFDM is currently a very popular choice for
future wireless applications, including
wireless LANs, cellular and PCS data, and
possibly Fourth Generation systems.
SUMMARY
• Introduction
• Basic Concepts
• PeaktoAverage Power Ratio Reduction
• Time and Frequency Synchronization
• Channel Estimation
• Applications
• Summary
• References
OUTLINE
1. M.L. Doelz, E.T. Heald, and D.L. Martin, “Binary data
transmission techniques for linear systems,” Proc. IRE,
pp. 656661, May 1957.
2. G.A. Franco and G. Lachs, “An orthogonal coding
technique for communications,” 1961 IRE International
Convention Record, pp. 126130.
3. J. Holsinger, “Digital communication over fixed time
continuous channel with memory  with special application
to telephone channels,” Ph.D. dissertation, Massachusetts
Institute of Technology, 1964.
4. P.A. Bello, “Selective fading limitations of the Kathryn
modem and some system design considerations,” IEEE
Trans. on Comm. Tech., pp. 320333, Sept. 1965.
5. R.W. Chang, “Synthesis of bandlimited orthogonal signals
for multichannel data,” pp. 17751797, BSTJ, Dec. 1966.
6. R.W. Chang, “Orthogonal frequency division
multiplexing.” U.S. Patent 3,488,445, filed Nov. 4 1966,
issued Jan. 1970.
7. M.S. Zimmerman and A.L. Kirsh, “The AN/GSC10
(KATHRYN) variable rate data modem for HF radio,” IEEE
Trans. on Comm. Tech., pp. 197205, April 1967.
8. B.R. Saltzberg, Performance of an efficient parallel data
transmission systems,” IEEE Trans. on Comm. Tech.
pp. 805811, Dec. 1967.
9. G.C. Porter, “Error distribution and diversity performance
of a frequencydifferential PSK HF modem,” IEEE Trans.
on Comm. Tech., pp. 567575, Aug. 1968.
REFERENCES
• Basic Concepts
10. R.W. Chang and R.A. Gibby, “A theoretical study of
performance of an orthogonal multiplexing data
transmission scheme,” IEEE Trans. on Comm. Tech., pp.
529541, Aug. 1968.
11. S.B. Weinstein and P.M. Ebert, “Data transmission by
frequencydivision multiplexing using the discrete Fourier
transform,” IEEE Trans. on Comm. Tech., pp. 628634,
October 1971.
12. B. Hirosaki, “An analysis of automatic equalizers for
orthogonally multiplexed QAM systems,” IEEE Trans. on
Comm. Tech., pp. 7383, Jan. 1980.
13. A. Peled and A. Ruiz, “Frequency domain data
transmission using reduced computational complexity
algorithms,” Proc. of ICASSP ’80, pp. 964967.
14. W.E. Keasler, Jr., “Reliable data communications over the
voice bandwidth telephone using orthogonal frequency
division multiplexing.” Ph.D. dissertation, Univ. of Illinois,
UrbanaChampaign, 1982.
15. L.J. Cimini, Jr., “Analysis and simulation of a digital
mobile channel using orthogonal frequency division
multiplexing,” IEEE Trans. on Comm., pp. 665675, July
1985.
16. B. Hirosaki, S. Hasegawa, and A. Sabato, “Advanced
groupband data modem using orthogonally multiplexed
QAM techniques,” IEEE Trans. on Comm., pp. 587592,
June 1986.
17. I. Kalet, “The multitone channel,” IEEE Trans. on Comm.,
pp. 119124, Feb. 1989.
REFERENCES
• Basic Concepts (cont’d)
18. J.A.C. Bingham, ”Multicarrier modulation for data
transmission: An idea whose time has come,” IEEE
Comm. Mag., pp. 514, May 1990.
19. E.F. Casas and C. Leung, “OFDM for data communication
over mobile radio FM channels,” IEEE Trans. on Comm.,
pp. 783793, May 1991.
20. Special Issue on MultiCarrier Modulation, Kluwer Wireless
Pers. Comm., January 1996.
21. L.J. Cimini, Jr., and N.R. Sollenberger, “OFDM with
diversity and coding for highbitrate mobile data
applications,” Proc. of the 3rd International Workshop on
Mobile Multimedia Comm., Sept. 1996.
22. L.J. Cimini, Jr., B. Daneshrad, and N.R. Sollenberger,
“Clustered OFDM with transmitter diversity and coding,”
Proc. of Globecom ‘96, pp. 703707.
REFERENCES
• Basic Concepts (cont’d)
1. R.S. Kennedy, Fading Dispersive Communication
Channels, Wiley, 1969.
2. W.C. Jakes, Jr., Ed., Microwave Mobile Communications,
Wiley, 1974.
3. J.D. Parsons, The Mobile Radio Propagation Channel,
Wiley, 1992.
4. J.B. Anderson, T.S. Rappaport, and S. Yoshida,
“Propagation measurements and models for wireless
communications channels,” IEEE Comm. Mag., pp. 4249,
January 1995.
• Radio Environment
5. R. Steele, Mobile Radio Communications, IEEE Press,
1995.
6. T.S. Rappaport, Wireless Communications, IEEE Press/
PrenticeHall, 1996.
REFERENCES
• Radio Environment (cont’d)
1. A.E. Jones, T.A. Wilkinson, and S.K. Barton, “Block coding
scheme for reduction of peak to mean envelope power
ratio of multicarrier transmission schemes,” Elect. Lett.,
pp. 20982099, Dec. 1994.
2. R. O’Neil and L.B. Lopes, “Envelope variations and
spectral splatter in clipped multicarrier signals,” Proc. of
PIMRC ‘95, pp. 7175.
3. P. Van Eetvelt, G. Wade, and M. Tomlinson, “Peak to
average power reduction for OFDM schemes by selective
scrambling,” Elect. Lett., pp. 19631964, Oct. 1996.
4. R.W. Bäuml, R.F.H. Fischer, and J.B. Huber, “Reducing
peaktoaverage power ratio of multicarrier modulation by
selected mapping.” Elect. Lett., pp. 20562057, Oct. 1996.
5. R.D.J. van Nee, “OFDM codes for peaktoaverage power
reduction and error correction,” Proc. of Globecom ‘96,
pp. 740744.
6. S.H. Müller and J.B. Huber, “OFDM with reduced peakto
average power ratio by optimum combination of partial
transmit sequences,” Elect. Lett., pp. 368369, Feb. 1997.
7. X. Li and L.J. Cimini, Jr., “Effects of clipping and filtering
on the performance of OFDM,” IEEE Comm. Letts.,, pp.
131133, May 1998.
• PeaktoAverage Power Reduction
8. S.H. Müller and J.B. Huber, “A novel peak power reduction
scheme for OFDM,” Proc. of PIMRC ‘97, pp. 10901094.
9. M. Friese, “Multitone signals with low crestfactor,” IEEE
Trans. on Comm., pp. 13381344, Oct. 1997.
10. J. Tellado and J.M. Cioffi, “PAR reduction in multicarrier
transmission systems,” ANSI T1E1.4 committee
contribution, number 97367, Dec. 1997.
REFERENCES
• PeaktoAverage Power Reduction (cont’d)
1. P.H. Moose, “A technique for orthogonal frequency
division multiplexing frequency offset correction,” IEEE
Trans., on Comm., pp. 29082914, Oct. 1994.
2. T. Pollet, P. Spruyt, and M. Moeneclaey, The BER
performance of OFDM systems using nonsynchronized
sampling,” Proc. of Globecom ‘94, pp. 253257.
3. K.W. Kang, J. Ann, and H.S. Lee, “Decisiondirected
maximumlikelihood estimation of OFDM frame
synchronization offset,” Elect Lett., pp. 21532154, Dec.
1994.
4. J.S. Oh, Y.M. Chung, and S.U. Lee, “A carrier
synchronization technique for OFDM on the frequency
selective fading environment,” Proc. of VTC ‘96, pp. 1574
1578.
5. F. Daffara and O. Adami, “A novel carrier recovery
technique for orthogonal multicarrier systems,” European
Trans. on Telecomm., pp. 323334, JulyAug. 1996.
• Time and Frequency Synchronization
6. H. Nogami and T. Nagashima, “A frequency and timing
period acquisition technique for OFDM systems,” IEICE
Trans. on Comm., pp. 11351146, Aug. 1996.
7. M. Luise and R. Reggiannini, “Carrier frequency
acquisition and tracking for OFDM systems,” IEEE Trans.
on Comm., pp. 15901598, Nov. 1996.
8. M. Speth, F. Classen, and H. Meyr, “Frame synchronization
OFDM systems in frequency selective fading channels,”
Proc. of VTC ‘97, pp. 18071881.
9. L. Hazy and M. ElTanany, “Synchronization of OFDM
systems over frequency selective fading channels,” Proc.
of VTC ‘97, pp. 20942098.
10. J.J. van de Beek, M. Sandell, and P.O. Börjesson, “ML
estimation of timing and frequency offset in OFDM
systems,” IEEE Trans. of Sig. Proc., pp. 18001805, July
1997.
11. D. Lee and K. Cheun, “A new symbol timing recovery
algorithm for OFDM systems,” IEEE Trans. on Consum.
Elect., pp. 767775, Aug. 1997.
12. T. Schmidl and D. Cox, “Robust frequency and timing
synchronization for OFDM,” IEEE Trans. on Comm., pp.
16131621, Dec. 1997.
REFERENCES
• Time and Frequency Synchronization (cont’d)
1. J.J. van de Beek, O. Edfors, M. Sandell, S.K. Wilson,
and P.O. Börjesson, “On channel estimation in OFDM
systems,” Proc. of VTC ‘95, pp. 815819.
• Channel Estimation
2. V. Mignone and A. Morello, “CD3OFDM: a novel
demodulation scheme for fixed and mobile receivers,”
IEEE Trans. of Comm., pp. 11441151, Sept. 1996.
3. P. Hoeher, S. Kaiser, and P. Robertson, “Twodimentional
pilotsymbolaided channel estimation by Wiener filtering,”
Proc. of ICASSP ‘97, pp. 18451848.
4. H.H. H’mimy, “Channel estimation based on coded pilot
for OFDM,” Proc. of VTC ‘97, pp. 13751379.
5. F. Tufvesson and T. Maseng, “Pilot assisted channel
estimation for OFDM in mobile cellular systems,” Proc. of
VTC ‘97, pp. 16391643.
6. P. Hoeher, S. Kaiser, and P. Robertson, “Pilotsymbol
aided channel estimation in time and frequency,” Proc. of
Globecom ‘97, pp. 9096.
7. Y.(G.) Li, L.J. Cimini, Jr., and N.R. Sollenberger,
“Robust channel estimation for OFDM systems with rapid
dispersive fading channels,” IEEE Trans. on Comm., pp.
902915, July 1998.
8. O. Edfors, M. Sandell, J.J. van de Beek, S.K. Wilson, and
P.O. Börjesson, “OFDM channel estimation by singular
value decomposition,” IEEE Trans. on Comm., pp. 931939,
July 1998.
9. Y.(G.) Li and N. Sollenberger, “Interference
suppression in OFDM systems using adaptive antenna
arrays,” IEEE Trans. on Comm., pp. 217229, Feb. 1999.
10. Y.(G.) Li, N. Seshadri, and S. Ariyavisitakul,
Transmitter diversity of OFDM systems with dispersive
fading channels,” IEEE JSAC, pp.461471, March 1999.
REFERENCES
• Channel Estimation (cont’d)
1. M. Alard and R. Lassalle, “Principles of modulation and channel
coding for digital broadcasting for mobile receivers,” EBU
Review  Technical, Aug. 1987.
2. B. Le Floch, R. HalbertLassalle, and D. Castelain, “Digital
sound broadcasting to mobile receivers,” IEEE Trans. of
Broad., pp. 493503, Aug. 1989.
3. D. Raychaudhuri and N.D. Wilson, “ATMbased transport
architecture for multiservices wireless personal
communication networks,” IEEE JSAC, pp. 14011413,
Oct. 1992.
4. “Radio broadcasting systems: Digital audio broadcasting
(DAB) to mobile, portable and fixed receivers,” ETS 300
401, ETSI  European Telecommunications Standards
Institute, Valbonne, France, Feb. 1995.
5. IEEE P802.11D3, Wireless LAN Medium Access Control
(MAC) and Physical Layer (PHY) Specifications, IEEE
Standard Department, Jan. 1996.
6. K.M. Aldinger, “A Multicarrier scheme for HIPERLAN.”
Wireless Personal Comm., Kluwer Acad. Pub., Jan. 1997.
7. L.J. Cimini, Jr., J.C. Chaung, and N.R. Sollenberger,
“Advanced Cellular Internet Service (ACIS),” IEEE Comm.
Mag., pp. 150159, Oct. 1998.
REFERENCES
• Applications
ORTHOGONAL FREQUENCY DIVISION MULTIPLEXING FOR WIRELESS CHANNELS
Leonard J. Cimini, Jr.
AT&T Labs – Research 100 Schulz Dr., Rm. 4146 Red Bank, NJ 077017033, USA Email: ljc@research.att.com Tel/Fax: 17323453129/3039
Ye (Geoffrey) Li
AT&T Labs – Research 100 Schulz Dr., Rm. 4152 Red Bank, NJ 077017033, USA Email: ljye@research.att.com Tel/Fax: 17323453132/3039
ABSTRACT
Orthogonal frequency division multiplexing (OFDM) has been shown to be an effective technique to combat multipath fading in wireless communications. It has been successfully used for HF radio applications and has been chosen as the standard for digital audio broadcasting and digital terrestrial TV broadcasting in Europe and highspeed wireless local areas networks. In this tutorial, we present the basic principles of OFDM and discuss the problems, and some of the potential solutions, in implementing an OFDM system. Techniques for peaktoaverage power ratio reduction, time and frequency synchronization, and channel estimation will be discussed. We conclude with a brief overview of current application areas
BIOGRAPHIES OF PRESENTERS:
Leonard J. Cimini, Jr., received the B.S.E. (summa cum laude), M.S.E. and Ph.D. degrees in electrical engineering from the University of Pennsylvania in 1978, 1979, and 1982, respectively. During the graduate work he was supported by a National Science Foundation Fellowship. Since 1982, he has been employed at AT&T, where his research interests are in wireless communications systems. Dr. Cimini is a member of Tau Beta Pi and Eta Kappa Nu. He has been very active in the IEEE Communications Society and is EditorinChief of the IEEE JSAC: Wireless Communications Series. He is also an Adjunct Professor at the University of Pennsylvania. Ye (Geoffrey) Li received the B.Eng and M.Eng degrees in 1983 and 1986, respectively, from the Department of Wireless Engineering, Nanjing Institute of Technology, Nanjing, China, and the Ph.D. degree in Electrical Engineering in 1994, Auburn University, Alabama. Since May 1996, he has been with AT&T Labs  Research. His current research interests are in statistical signal processing and wireless communications. He has served as a guest editor for a special issue on Signal Processing for Wireless Communications for the IEEE JSAC and is an editor for Wireless Communication Theory for the IEEE Transactions on Communications.
OUTLINE
• Introduction • Basic Concepts • PeaktoAverage Power Ratio Reduction • Time and Frequency Synchronization • Channel Estimation • Applications • Summary • References
OUTLINE • Introduction • Basic Concepts • PeaktoAverage Power Ratio Reduction • Time and Frequency Synchronization • Channel Estimation • Applications • Summary • References .
INTRODUCTION • Motivation • Radio Environment • Brief History .
MOTIVATION • Highbitrate wireless applications • Limitations caused by the radio environment • OFDM can overcome these inherent bit rate limitations .
PATH LOSS MODEL • Path Loss • Shadow Fading • Multipath • Flat fading • Doppler spread • Delay spread • Interference .
Pt is the transmitted power. The path loss exponent α = 2 in free space. is L= Pr 1 =K α Pt d where Pr is the local mean received signal power. . and d is the transmitter receiver distance. models are used for different environments. • A simple model for path loss. often complicated.PATH LOSS MODEL • Different. L. 2 ≤ α ≤ 4 in typical environments.
σ 2 . • This results in variations in the local mean received signal power.SHADOW FADING • The received signal is shadowed by obstructions such as hills and buildings. S ( ) • Implications – nonuniform coverage – increases the required transmit power . 4 ≤ σ S ≤ 10 dB. Pr (dB ) = Pr (dB ) + GS where GS ~ N 0.
in seconds 20 x.5λ 0 0 0.5 10 1 t. in wavelength 1.MULTIPATH Received Power Delay Spread t h(t ) = ∑ aie jθi δ(t − t i ) i Constructive and destructive interference of arriving rays 10 0 dB With Respect 10 to RMS Value 20 30 0.5 30 .
follows a Rayleigh or Rician distribution. r. • The received signal envelope. P (dB) = P (dB) + G + 20 log r r r S Received Signal Power (dB) path loss shadow fading Rayleigh fading log (distance) • Implications – increases the required transmit power – causes bursts of errors .FLAT FADING • The delay spread is small compared to the symbol period.
5 mph. fD = 37 Hz • Implications – signal amplitude and phase decorrelate after a time period ~ 1/fD .DOPPLER SPREAD • A measure of the spectral broadening caused by the channel time variation. fD ≤ v λ Example: 900 MHz. 60 mph. fD = 80 Hz 5 GHz.
DELAY SPREAD TIME DOMAIN INTERPRETATION Received Power 2τ Tworay model τ = rms delay spread Delay Channel Input τ small T 0 2T Channel Output 1 0 T 1 T 2T T τ large 0 T 2T τ • T small • τ large T negligible intersymbol interference significant intersymbol interference. which causes an irreducible error floor .
DELAY SPREAD FREQUENCY DOMAIN INTERPRETATION H(f) Bs = signal bandwidth ≈ 1/T Bs 1 2τ f τ • T small flat fading frequencyselective fading • τ T large .
5 µsec 500 nsec 100 nsec τ Maximum Bit Rate 8 kbps 80 kbps 400 kbps 2 Mbps .BIT RATE LIMITATIONS • ISI causes an irreducible error floor. for QPSK Mobile (rural) Mobile (city) Microcells Large Building 25 µsec 2. For example. 101 Coherent Detection + BPSK QPSK OQPSK Modulation x MSK Irreducible Pb 102 x x x x + + + + x + 103 104 102 101 rms delay spread τ = symbol period T 100 The rms delay spread imposes a limit on the maximum bit rate.
S S 1D ≈ = I+N I 6R α • Implications – high reuse efficiency requires interference mitigation .INTERFRENCE • Frequencies are reused often to maximize spectral efficiency. the noise floor is dominated by cochannel interference. R BASE STATION D • For interferencelimited systems.
Gandalf) • Cellular modem (Telebit) • Digital audio and terrestrial TV broadcasting (Europe) • Asymmetric digital subscriber line (DMT) • Wireless LANs – IEEE802.HISTORY • Military HF radio (1950’s .1960’s) – Kineplex – Kathryn • Wireline modem (Telebit.11 .National Information Infrastructure – HIPERLAN TYPE II .
OUTLINE • Introduction • Basic Concepts • PeaktoAverage Power Ratio Reduction • Time and Frequency Synchronization • Channel Estimation • Applications • Summary • References .
BASIC CONCEPTS • Multicarrier • Basic OFDM • Impairments • Alternative forms .
• Ideally. each subchannel is narrow enough so that the fading it experiences is flat ⇒ no ISI. d0(t) d1(t) d N1(t) f0 RF f1 fN1 D(t) Transmitter R/N b/s R/N b/s R/N b/s QAM QAM QAM filter filter filter Bandlimited signals f0 f1 filter f0 filter f1 f2 QAM f0 QAM f1 filter fN1 Receiver RF QAM fN1 • Disadvantage: .MULTICARRIER • The transmission bandwidth is divided into many narrow subchannels which are transmitted in parallel.Spectrally inefficient .Requires filter bank at receiver .
Horizontal slide here .
then d(m) = d(m) – choose fn = f0 + n∆f. . interblock interference) ⇒ lower spectral or power efficiency.BASIC OFDM RECEIVER ∫ f0 RF d(0) d(1) parallel to serial converter ∫ f1 QAM ∫ f N1 d(N1) • Subchannel separation 1 NT ^ – integrate over NT. with ∆f = ∆ • A guard interval can virtually eliminate ISI (or.
0 ≤ t ≤ Ts k = 0 • Baseband x b (t ) = ∑ a[k ]e j2 πk∆ft .PASSBAND VERSUS BASEBAND • Passband N− 1 j2 π ( fc +k∆f ) t x p (t ) = ℜ ∑ a[k ]e . 0 ≤ t ≤ Ts k =0 N −1 .
∆fTs = 1. n N−1 x b [n] = x b TS = ∑ a[k ]e j2 πnk∆fTs / N N k =0 • For orthogonality. 0 ≤ t ≤ Ts k =0 N−1 • If sampled at a rate of Ts /N.DFT IMPLEMENTATION TRANSMITTER • Transmitted signal can be obtained using a Discrete Fourier Transform x b (t ) = ∑ a[k ]e j2 πk∆ft . x b [n] = ∑ a[k ]e j2 πnk / N = IDFT{a[k ]} k =0 N− 1 • Efficient FFT implementation .
DFT IMPLEMENTATION RECEIVER ˆ a[k ] = DFT{x b [n]} 1 N −1 = ∑ x b [n] e − j2 πnk / N N n=0 1 N −1 N − 1 = ∑ ∑ a[m] e j2 πn(m−k ) / N N n = 0 m= 0 N−1 1 N− 1 = ∑ a[m] ∑ e j2 πn(m−k ) / N N m=0 n= 0 1 N− 1 = ∑ a[m] N δ[m − k ] N m=0 = a[k ] .
PERFORMANCE IMPROVEMENT • Coding across subchannels ⇒ works best with large delay spread • Adaptive loading – More bits/symbol where SNR is sufficient – Could also adapt transmit power in each subchannel – Requires reliable feedback channel and accurate channel information • Frequency equalization and coherent detection ⇒ requires accurate channel estimation .
2 Mbits/sec using QPSK with B=800 kHz bandwidth channel – Delay span up to 40 µsec (max 5 kbaud for single carrier) • Design – Choose subchannel width so that there is no ISI in each subchannel ⇒ ∆ f = 6. each transmitting 2 bits in 200 µsec Rb = 120 x 2 bits = 1.SAMPLE DESIGN • Goal – Transmit 1.25 kHz ⇒ N = B/∆f = 128 subchannels ∆ – OFDM symbol duration Ts=1/∆f = 160 µsec ∆ – Guard interval Tg = 40 µsec – OFDM block length: Tf = Ts + Tg = 200 µsec – Assuming 4 guard channels on each end. there are 120 data subchannels.2 Mbits / sec 200 µ sec .
IMPAIRMENTS • Timevarying fading. and timing mismatch impair the orthogonality of the subchannels. • Large amplitude fluctuations can be a serious problem when transmitting through a nonlinearity. frequency offset. .
k ] LO O M N • Frequency offset – For a frequency offset between the transmitter and receiver. k ] n≠k LO MOO O N attenuated & rotated ICI a[k ]K[k.TIMEVARYING IMPAIRMENTS • General expression: ˆ a[k ] = DFT{χ b [n]} = + ∑ a[n]K[n. k ] = π δf ∆f . 0 ≤ t ≤ Ts k =0 δf sin π n − k − jπ n−k − δf ∆f ∆f K[n. N− 1 x b (t ) = ∑ a[k ] e j2 π (k∆f − δf )t . k ] = e δf π n − k − ∆f δf sin π jπ δf ∆f e ∆f K[k.
. ˆ a δt [k ] = a[k ]e j2 πfk δt ⇓ a phase shift • Otherwise.TIMING MISMATCH • Timing offset smaller than the guard interval results in a phase shift. additional interference is generated. • Best solution is to choose sufficient guard interval.
DELAY SPREAD • Assuming timeinvariance. the multipath channel results in a received signal at kth subchannel ˆ a[k ] = H[k ]a[k ] + w [k ] • Simple complex multiplicative distortion • For coherent detection. . channel parameter estimation and tracking are required.
NONLINEARITIES • Large peaktoaverage power ratio (PAPR) • PAPR ~ number of subcarriers • Large PAPR ⇒ inband distortion and spectral spreading • PAPR reduction techniques required .
1966) – B..W. 1980) • WaveletBased OFDM: – B.R. on Comm.. 1967) – B. Jan. IEEE.ALTERNATIVE FORMS • Bandlimited OFDM: – R. LeFloch (Proc. on Comm. Dec. Hirosaki (IEEE Trans. Saltzberg (IEEE Trans. Chang (BSTJ. Dec. June 1995) . Tech.
OUTLINE • Introduction • Basic Concepts • PeaktoAverage Power Ratio Reduction • Time and Frequency Synchronization • Channel Estimation • Applications • Summary • References .
PAP RATIO REDUCTION • PAPR properties • Clipping and filtering • Selective mapping • Partial transmit sequences • Coding • Other techniques .
PAPR PROPERTIES
• Superposition of a large number of subcarrier
signals results in a Rayleigh envelope.
• PAPR definition
PAPR = max 0< t < Ts x b (t ) Pave
2
=N
For N=128, PAPR = 21 dB.
• However, these large peaks do not occur very
often.
PAPR PROPERTIES
• Large PAPR – In band noise ⇒ increases BER – Spectral spreading ⇒ ACI • Possible Solutions – Amplifier backoff – Reduce PAPR of OFDM signal
CLIPPING AND FILTERING
• Deliberate clipping will reduce peak value,
but will result in spectral spreading (ACI) and inband distortion (BER).
• Filtering is required to minimize spectral
spreading. ⇒ peak regrowth
CLIPPING AND FILTERING
100
101
BER
102
CR=0.8 1.0 1.2
103
Unclipped
1.4
1.6 104 0 2 4 6 8 10 12 14 16
SNR (dB)
0 10 20 30 CR = 1.4
PSD (dB)
40 50 60 70 80 90 10 4 3 2 1 0 1 2 3 4
NORMALIZED FREQUENCY
• Convert each data sequence into the timedomain with an Npoint IFFT • Select sequence for transmission with the smallest PAPR . r1.… rM.SELECTIVE MAPPING • Multiply data signal by M different sequences.
PARTIAL TRANSMIT SEQUENCE • Divide the OFDM tones into M clusters • Convert each cluster into the timedomain using an Npoint IFFT • Combine the M output sequences to minimize the PAPR .
PTS can achieve a lower PAPR at the expense of more complexity.PERFORMANCE Using the same redundancy. .
CODING • Nonlinear – Map transmitted sequence into a larger sequence where highpeak sequences are not used – Good performance with little overhead – Requires table lookup ⇒ only applicable for small number of subchannels – Error propagation • Current work searching for systematic implementation with some error correction capability .
OUTLINE • Introduction • Basic Concepts • PeaktoAverage Power Ratio Reduction • Time and Frequency Synchronization • Channel Estimation • Applications • Summary • References .
TIME AND FREQUENCY SYNCHRONIZATION • Timing offset estimation • Frequency offset estimation • Joint offset estimation .
TIMING OFFSET ESTIMATION • Pilotbased methods – NonOFDMbased pilot symbols – OFDMbased pilot symbols • Nonpilot based methods .
PILOTBASED TIMING NonOFDM Pilot Symbols Use a null signal inserted at the start of each group of OFDM blocks OFDM OFDM Null Symbol OFDM OFDM ... OFDM Null Symbol time OFDM .
frequency Cyclic prefix First Half of Symbol = Second Half of Symbol time .PILOTBASED TIMING OFDM Pilot Symbols Design special OFDM block for estimation (Moose.. Schmidl) N tones ..
35 0.4 Schmidl estimator 0.25 0.PILOTBASED TIMING Performance 0.45 'ML' pilot timeoffset estimator probability timeoffset estimator outside saleregion of pilotregion 0.15 0.05 0 2 4 6 8 10 12 14 16 18 20 SNR (dB) .3 0.2 0.1 0.
06 0.03 0.07 0.01 0 2 4 6 8 10 12 14 16 18 20 SNR (dB) .08 0.02 0.09 probability timeoffset estimator outside sale region 0.05 0.04 0.NONPILOT BASED TIMING Use redundancy in the cyclic prefix to estimate the time offset 0.1 ML Squareddifference Correlation only 0.
FREQUENCY OFFSET ESTIMATION • Pilotbased methods – NonOFDMbased pilot symbols – OFDMbased pilot symbols • Nonpilot based methods .
...FREQUENCY OFFSET ESTIMATION PilotBased An OFDMbased pilot scheme for coarse and fine frequency synchronization OFDM symbol with pilots OFDM symbol . . Pilot tone .. OFDM symbol OFDM symbol with pilots time D1 OFDM symbols Frequency domain of an OFDM pilot symbol Pilot tones spaced every 4 tones.
15 0.05 0 2 4 6 8 10 12 14 16 18 20 SNR (dB) .1 epsilon = .3 0.FREQUENCY OFFSET ESTIMATION Performance 0.3 probability of incorrect toneordering 0.25 0.35 epsilon = 0 epsilon = .2 0.2 epsilon = .1 0.
4 0.8 0.6 1.4 1.2 1 0.JOINT ESTIMATION Based on shortened Moose pilot symbol 8 standard deviation of tim eoffset estim ator 7 6 5 4 3 'ML' pilotbased estimator Schmidl estimator 2 1 2 4 6 8 10 12 14 16 18 20 SNR (dB) 1.6 0.2 0 2 4 6 8 10 12 14 16 18 20 SNR (dB) 'ML' pilotbased estimator Schmidl estimator Estimated variance .
03 0.06 Joint ML Squareddifference Shortened squared difference Estimated variance 0.JOINT ESTIMATION Based on a 16sample cyclic prefix 30 Joint ML Squareddifference Shortened squared difference 20 25 15 10 5 0 2 4 6 8 10 12 14 16 18 20 SNR (dB) 0.05 0.02 0.04 0.01 0 2 4 6 8 10 12 14 16 18 20 SNR (dB) .
OUTLINE • Introduction • Basic Concepts • PeaktoAverage Power Ratio Reduction • Time and Frequency Synchronization • Channel Estimation • Applications • Summary • References .
CHANNEL ESTIMATION • Differential and coherent detection • Decisiondirected estimation • Pilotsymbolaided estimation .
DIFFERENTIAL AND COHERENT DETECTION • 3dB SNR degradation for differential detection • Differential – in time ⇒ sensitive to Doppler shift – in frequency ⇒ sensitive to delay spread • Coherent detection requires channel information .
DECISIONDIRECTED ESTIMATION • Use sliced data for estimating channel parameters • Obtain an MMSE or robust estimator using the correlations of the channel parameters in time and/or frequency. .
K] ^ H[n. . . U . .DECISIONDIRECTED ESTIMATION MMSE Estimator • Using time and frequency correlations • Estimator coefficients C[n] = UΦ[n]UH Φ[n] = diag{Φ 1[n]. . . . ω Φ1(ω) Φk0(ω) ω . 2] ∼ H[n. .. 1] ∼ H[n. . . . 2] . . . . . . K] . . .. . 1] ^ H[n. . . Φ N [n]} • Estimator structure ∼ H[n. ^ H[n. . UH . . . . . . . . . . . . . . . . .
• Robust design – Match rectangular spectrum in time and frequency domains – Good performance for almost all channels – Relatively insensitive to Doppler and delay profiles .DECISIONDIRECTED ESTIMATION Robust Estimator • Why robust estimation? – A large performance degradation is possible if MMSE estimator is not matched to the channel.
PILOTSYMBOLAIDED ESTIMATION • Pilot symbol grid • Obtain an estimate of the channel at the pilot symbol positions. • Obtain estimates at other frequencies and times by interpolation. .
PILOTSYMBOLAIDED ESTIMATION Grid Selection • 2D spectrum of channel parameters at pilot symbol positions • Nonrectangular pilot symbol grids are better .
which corrects 10 erasures.SIMULATION PARAMETERS • 800 kHz bandwidth • Number of subchannels N = 128 subchannels (4 guard subchannels at each end) • OFDM block duration Tf = 200 µsec (with 40µsec guard interva) • (40.20) RS code. based on signal strength. and correct 5 random errors .
WORD ERROR RATE Typical Urban Channel .
WORD ERROR RATE Hilly Terrain Channel .
OUTLINE • Introduction • Basic Concepts • PeaktoAverage Power Ratio Reduction • Time and Frequency Synchronization • Channel Estimation • Applications • Summary • References .
APPLICATIONS • Asymmetric digital subscriber line • Digital audio and terrestrial TV broadcasting • Wireless LAN’s • Highspeed cellular data .
DIGITAL AUDIO BROADCASTING • Broadcasting standard in Europe • Single frequency network • Bandwidth = 7 MHz • Useful bit rate = 5. coding with constraint length 7 • Time interleaving: 16 X 24 msec • Pilot symbols for channel estimation .6 Mbits/sec • N = 448 subchannels • Tf = 80 µsec. and Tg = 16 µsec • Rate1/4 conv. Tf = 64 µsec.
16 QAM in each subchannel 48 Subchannels out of 64 Convolution K=7. R=1/2 or R=3/4 intercarrier interleave 5 Mbps (BPSK. R=3/4) . R=1/2) 10 Mbps (QPSK. R=1/2) 15 Mbps (QPSK. R=1/2) 30 Mbps (16QAM.WIRELESS LANS • IEEE802. R=3/4)) 20 Mbps (16QAM.11(a) – New NII spectrum at about 5 GHz – Indoor applications Modulation scheme Subcarrier Modulation Number of subchannels Coding Data rates OFDM DBPSK.
2 GHz – HIPERLAN Type I: single carrier GSMK with equalization – HIPERLAN Type II: COFDM Information data rate Modulation Coding rate Coding OFDM symbol duration Guard interval Tprefix Tpostfix Subchannel spacing Rolloff factor Channel Spacing Occupied 3 dB Bandwidth 48. 32.666 kHz 0.025 25 MHz 20 MHz . QPSK or BPSK Convolutional 3/4 or 1/2 48 3 µs 600 ns 600 ns 75 ns 416.WIRELESS LANS • HIPERLAN: High Performance Radio Local Area Network – Standard in Europe at 5. 24. 16 and 8 Mbps OFDM with 16QAM.
8 µsec – ramp time: Tr = 10 µsec – block length: Tf = Ts+ Tr + Tg = 288.5 µsec – cluster width: Bc = Nc ∆f = 100 kHz .17 kHz – symbol duration: Ts = 1/∆f = 240 µsec ∆ – guard interval: Tg = 38. Japan • Combination of TDMA. OFDM and cluster hopping • Parameters – number of tones per cluster: Nc = 24 – tone spacing: ∆f = 4.HIGH SPEED CELLULAR Band Division Multiple Access (BDMA) • Proposal by Sony.
5 µsec – rate1/2 RS code across subchannels .17 kHz – symbol duration: Ts = 1/∆f = 240 µsec ∆ – guard interval: Tg = 48.HIGH SPEED CELLULAR Advanced Cellular Internet Service (ACIS) • Goal to provide widearea Internet service to mobile subscribers • Combines OFDM with multiple transmitter and receiver antennas and coding • Parameters – total bandwidth: B = 800 kHz – number of tones: N = 192 – tone spacing: ∆f = B/N = 4.5 µsec – block length: Tf = Ts + Tg = 288.
OUTLINE • Introduction • Basic Concepts • PeaktoAverage Power Ratio Reduction • Time and Frequency Synchronization • Channel Estimation • Applications • Summary • References .
to realize an OFDM system. several practical issues must be addressed.SUMMARY • Highbitrate wireless data is desirable. but the radio environment puts an upper limit on the achievable bit rate. • Several promising solutions have been proposed for all of these problems. • OFDM is currently a very popular choice for future wireless applications. by transmitting data over many narrow subchannels. • However. cellular and PCS data. can overcome the bit rate limit. and possibly Fourth Generation systems. including wireless LANs. . and channel estimation. including PAPR. frequency offset and timing mismatch. • OFDM.
OUTLINE • Introduction • Basic Concepts • PeaktoAverage Power Ratio Reduction • Time and Frequency Synchronization • Channel Estimation • Applications • Summary • References .
Kirsh. “Digital communication over fixed timecontinuous channel with memory . 4 1966. 126130. 197205.S. Porter. 4. Tech. 6.” IEEE Trans. 320333. 1968. “Binary data transmission techniques for linear systems. Dec.” IEEE Trans. “The AN/GSC10 (KATHRYN) variable rate data modem for HF radio. on Comm. BSTJ. M. 7. 5.W. 17751797.L. B. “Synthesis of bandlimited orthogonal signals for multichannel data. Chang.with special application to telephone channels. E. May 1957. Sept.REFERENCES • Basic Concepts 1. April 1967. 3. R. M. IRE. 567575.A. 2. pp. G.. Holsinger. J. Doelz. Bello. P. “Selective fading limitations of the Kathryn modem and some system design considerations. Dec. 805811.” pp.S.” 1961 IRE International Convention Record. .” IEEE Trans. Tech.C.T. Franco and G. Tech.L. Performance of an efficient parallel data transmission systems. 9.L. pp. Chang. Heald. G. on Comm. filed Nov.W..445. 8. issued Jan. on Comm. “Error distribution and diversity performance of a frequencydifferential PSK HF modem. pp. Martin. Aug. Patent 3. dissertation. Saltzberg.” IEEE Trans.488. 1965. “An orthogonal coding technique for communications. 656661.” Ph. and D. 1966. pp. Tech.. “Orthogonal frequency division multiplexing. 1967. pp. Zimmerman and A. Lachs.D.” U.R. on Comm. 1964.A.” Proc. R. Massachusetts Institute of Technology. 1970. pp.
pp. on Comm. pp. 628634.. “A theoretical study of performance of an orthogonal multiplexing data transmission scheme. 7383. pp. Hirosaki. “Reliable data communications over the voice bandwidth telephone using orthogonal frequency division multiplexing. “Frequency domain data transmission using reduced computational complexity algorithms. Tech.REFERENCES • Basic Concepts (cont’d) 10.E. UrbanaChampaign.. Hirosaki. Sabato. S.A. Cimini. 964967. June 1986. Tech. 119124. 11. R. Ruiz. W.D. S. 12. 16. “An analysis of automatic equalizers for orthogonally multiplexed QAM systems. pp. on Comm. B. Weinstein and P. 529541. Jan. Feb. I. 1989. B.” IEEE Trans.B. Keasler.. Tech. Ebert. on Comm. “Analysis and simulation of a digital mobile channel using orthogonal frequency division multiplexing.J. Jr.W.” Ph. dissertation. Jr. Kalet.. 17. “Data transmission by frequencydivision multiplexing using the discrete Fourier transform. 587592. October 1971. Hasegawa. 665675. Peled and A. and A. . Gibby. Univ. Chang and R. 15.” Proc. A. pp.” IEEE Trans. 1982. Aug. 1980.M.” IEEE Trans.” IEEE Trans. of ICASSP ’80. on Comm. 1968. L. 14.” IEEE Trans. July 1985..” IEEE Trans. “Advanced groupband data modem using orthogonally multiplexed QAM techniques. of Illinois. “The multitone channel. on Comm. pp. pp.. on Comm.. 13..
and N. and N. Sept.B. Jr. “Propagation measurements and models for wireless communications channels. 22. J.A. Cimini. R. 1969. Mag.. B. 703707.” Proc. 20. Wiley. Jr. Special Issue on MultiCarrier Modulation. pp. L.. W.” Proc. and S. 21. T. The Mobile Radio Propagation Channel. Yoshida. pp. Leung. 1996.. 1992. Jakes. January 1996. 4249.C. of the 3rd International Workshop on Mobile Multimedia Comm. 783793.J.” IEEE Comm. Comm.C. J. Ed.” IEEE Trans. Fading Dispersive Communication Channels. pp. Cimini.REFERENCES • Basic Concepts (cont’d) 18. ”Multicarrier modulation for data transmission: An idea whose time has come. 19.. Mag.. May 1990. 3. 514. • Radio Environment 1. Sollenberger. 1974.. of Globecom ‘96. Bingham. May 1991.D. Sollenberger.J.. Casas and C.R. Daneshrad. Microwave Mobile Communications.. J. Parsons. . “OFDM with diversity and coding for highbitrate mobile data applications. E. Wiley. Anderson. 2. Jr. on Comm.F. January 1995. 4.R. L. “Clustered OFDM with transmitter diversity and coding. pp.S. Wiley. “OFDM for data communication over mobile radio FM channels.. Kennedy. Rappaport. Kluwer Wireless Pers.” IEEE Comm.S.
” Elect. and S. R. Mobile Radio Communications. Huber.. pp. of PIMRC ‘95. . R. May 1998.E. T. O’Neil and L. 6. Rappaport.W..” Elect. “OFDM with reduced peaktoaverage power ratio by optimum combination of partial transmit sequences. 6. Letts. 7. P.H.” IEEE Comm. Jr. pp. 1996. Lett.B. Lett.” Elect. 1995. 131133.K.F. 4. S. Wireless Communications. Wade. Oct. Cimini.H. 1997. and M. Huber. 368369. Lopes. “Effects of clipping and filtering on the performance of OFDM.S. Lett.J. IEEE Press/ PrenticeHall. • PeaktoAverage Power Reduction 1. of Globecom ‘96. R. 1994. 7175. Li and L. “Envelope variations and spectral splatter in clipped multicarrier signals... Barton. Dec.J.REFERENCES • Radio Environment (cont’d) 5. R.” Proc. “Block coding scheme for reduction of peak to mean envelope power ratio of multicarrier transmission schemes. and J. 20562057. “Peak to average power reduction for OFDM schemes by selective scrambling. 1996. Van Eetvelt. Oct. Feb. 740744. Steele. Tomlinson.” Elect. 1996. 5. Wilkinson.B. T.. Bäuml. R. “OFDM codes for peaktoaverage power reduction and error correction.A. X. 3.B. G. van Nee.. pp. 19631964. 2. pp. Müller and J. IEEE Press. pp. Jones. “Reducing peaktoaverage power ratio of multicarrier modulation by selected mapping. pp. Lett.” Proc. Fischer. 20982099.D.. A. pp.
1997. Huber.. Friese. Kang. Moose.” Proc. Oh.W. Lee.REFERENCES • PeaktoAverage Power Reduction (cont’d) 8. T. 1994. 10. Pollet. 323334. Dec. on Comm. JulyAug. and H. Lee.” ANSI T1E1. 1997. “A novel peak power reduction scheme for OFDM. pp.S. Oct. “A carrier synchronization technique for OFDM on the frequencyselective fading environment. Müller and J. F.” European Trans. “Multitone signals with low crestfactor. K. of VTC ‘96.M. 21532154. pp. on Telecomm. 1996. pp. “Decisiondirected maximumlikelihood estimation of OFDM frame synchronization offset. Dec. of Globecom ‘94. and M.” Proc.. Y. “PAR reduction in multicarrier transmission systems. S. Daffara and O. 253257. Tellado and J. J.. Moeneclaey. Oct. pp.” IEEE Trans. Ann. Cioffi.H. number 97367. M. J. on Comm.4 committee contribution. 3. 9.S. 13381344.U. • Time and Frequency Synchronization 1. pp. “A novel carrier recovery technique for orthogonal multicarrier systems.” Elect Lett. 2. 4. 29082914. 5. The BER performance of OFDM systems using nonsynchronized sampling. pp.” Proc.. and S. pp. of PIMRC ‘97. ..M. J. Spruyt.B. 15741578. 10901094. P.H. Adami. 1994.” IEEE Trans. Chung. “A technique for orthogonal frequencydivision multiplexing frequency offset correction. P.
1996. Hazy and M. Schmidl and D. July 1997. “A frequency and timing period acquisition technique for OFDM systems. Nagashima.. van de Beek.” IEEE Trans. Meyr. Aug.K. 15901598. 18001805. M. 7. on Comm. 767775. and P. pp.” IEEE Trans..REFERENCES • Time and Frequency Synchronization (cont’d) 6. Dec. 12. “Carrier frequency acquisition and tracking for OFDM systems. Speth. on Comm. S. pp. “On channel estimation in OFDM systems.” IEICE Trans.” IEEE Trans. “A new symbol timing recovery algorithm for OFDM systems. Reggiannini. 1997. 8.” IEEE Trans. 1996. H. 20942098.O. of VTC ‘97. and H. and P. “ML estimation of timing and frequency offset in OFDM systems. Börjesson. M. of Sig. on Consum. Proc. J. pp. Classen. 1997. L. Edfors. 9. “Frame synchronization OFDM systems in frequency selective fading channels. . F.. 18071881. Nov. Aug. J. of VTC ‘97. ElTanany. Sandell. “Synchronization of OFDM systems over frequency selective fading channels.” Proc.J. Elect. pp.” Proc. 16131621. 10. 11351146.O. O. 11. T. Sandell. on Comm.” Proc. D. “Robust frequency and timing synchronization for OFDM..J. Cox. M. pp. • Channel Estimation 1. M. Lee and K. pp. Börjesson.. Wilson. pp. Luise and R. Nogami and T. van de Beek. pp. Cheun. 815819. of VTC ‘95.
V. Y. pp.” Proc. of VTC ‘97. Kaiser.. Y. of VTC ‘97. F. J. July 1998. of ICASSP ‘97..” Proc. Wilson. and P. “Pilot assisted channel estimation for OFDM in mobile cellular systems. Robertson. Börjesson. S. Hoeher. 11441151. 8.(G. P. 9. N.” Proc.” IEEE Trans.) Li and N. and N. Sandell. 4. pp. pp. 931939. March 1999. Morello. 5. S. 10.) Li.REFERENCES • Channel Estimation (cont’d) 2.(G.H. pp. H’mimy. P. Kaiser. pp. 6. 3. “CD3OFDM: a novel demodulation scheme for fixed and mobile receivers. Seshadri. 13751379. Robertson. July 1998.J. and S.461471. 1996. Transmitter diversity of OFDM systems with dispersive fading channels.” IEEE JSAC. pp. 902915. Y.J. on Comm. 9096. “Channel estimation based on coded pilot for OFDM. O...” IEEE Trans. M. Feb. Maseng.” IEEE Trans. 7.) Li. Sollenberger. Sollenberger.” IEEE Trans. and P.O. 217229.” Proc. van de Beek. S.. Sept. Cimini. “Interference suppression in OFDM systems using adaptive antenna arrays. Tufvesson and T. Hoeher.K. pp. of Globecom ‘97. of Comm. “OFDM channel estimation by singular value decomposition. pp. Mignone and A. on Comm.(G. Ariyavisitakul. . pp. 18451848. H. and P. 1999. 16391643. “Robust channel estimation for OFDM systems with rapid dispersive fading channels. on Comm. Jr. “Pilotsymbolaided channel estimation in time and frequency. “Twodimentional pilotsymbolaided channel estimation by Wiener filtering. L. Edfors.R.
3. “Advanced Cellular Internet Service (ACIS). pp. Jan. R. HalbertLassalle.. Aug. J. “ATMbased transport architecture for multiservices wireless personal communication networks. Lassalle. 1987. 150159. Aug. 1992. ETSI . “Principles of modulation and channel coding for digital broadcasting for mobile receivers.Technical. M.J. L. IEEE P802. “Digital sound broadcasting to mobile receivers. 1998.” IEEE JSAC. and N. Le Floch. 2. of Broad. Jr. Alard and R. B. Jan. Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) Specifications. Castelain. . portable and fixed receivers.R. “Radio broadcasting systems: Digital audio broadcasting (DAB) to mobile.. K.” Wireless Personal Comm. pp.. France. Mag. Oct.European Telecommunications Standards Institute.” IEEE Comm..11D3.D. Wilson. 14011413.M. 493503. 4. “A Multicarrier scheme for HIPERLAN. 1989. IEEE Standard Department. Valbonne. 6. 1995. Kluwer Acad.C. 5. Cimini.” EBU Review . Oct. 1997. Sollenberger. Aldinger. Pub. D.REFERENCES • Applications 1. Feb.. 7. pp. Chaung. Raychaudhuri and N.” IEEE Trans. 1996. and D.” ETS 300 401.