Orthogonal Frequency

Division Multiplexing
OFDM
fred harris
Cubic Signal Processing Chair

San Diego State University

[email protected]
Vehicular Technology Conference - 2004
Textbooks and References
 “Wireless OFDM Systems: How to Make Them Work”
Marc Engels, Editor
 “OFDM Wireless LANs: A Theoretical and Practical Guide”
Juha Heiskala and John Terry
 “OFDM for Wireless Multimedia Communications”
Richard Van Nee and Ramjee Prasad
 “Single and Multi-Carrier Quadrature Amplitude Modulation”
Lajos Hanzo, William Webb, and Thomas Keller
 “ADSL, VDSL, and Multicarrier Modulation”
John Bingham
 “Implementing ADSL”
David Ginsburg
 “DSL Advances
Massimo Sorbara, John Cioffi, and Peter Silverman
OFDM
 OFDM also known as
Multi-Carrier or Multi-Tone Modulation
 DAB-OFDM
Digital Audio Broadcasting
 DVD-OFDM
Digital Video Broadcasting
 ADSL-OFDM
Asynchronous Digital Subscriber Line
 Wireless Local Area Network
IEEE-802.11a, IEEE-802.11g
ETSI BRAN (Hyperlan/2)
OFDM Systems
System Transform Number Channel Bandwidth Sample Symbol Data
Size Carriers Spacing MHz Rate Duration
kHz MHz sec Rate

Mbits/s
HyperLAN/2 64 52 312.5 16.25 20 3.2 6-54
4 0.8
802.11a 64 52 312.5 16.56 20 3.2 6-54
4 0.8

DVB-T 2048 1712 4.464 7.643 9.174 224 0.68-14.92

1024 842
DAB 2048 1536 1.00 1.536 2.048 24/48/96 3.072
8192 msec
ADSL 256 (down) 36-127 4.3125 1.104 1.104 231.9 0.64-8.192
64 (up) 7-28
 OFDM Advantages
 Efficiently Deals With Multi-path Fading
 Efficiently Deals With Channel Delay Spread
 Enhanced Channel Capacity
 Adaptively Modifies Modulation Density
 Robustness to Narrowband Interference
OFDM Disadvantages
 OFDM Sensitive to
Small Carrier Frequency Offsets
 OFDM Exhibits
High Peak to Average Power Ratio
 OFDM Sensitive to
High Frequency Phase Noise
 OFDM Sensitive to
Sampling Clock Offsets
Single Carrier System

Sequential Transmission
of Waveforms
Waveforms are
Short Duration T
Waveforms Occupy
Full System Bandwidth 1/T
Multi-Carrier System

Parallel Transmission
of Waveforms
Waveforms are
Long Duration MT
Waveforms Occupy 1/M th
Of System Bandwidth 1/T
OFDM: Dense Multichannel System

Conventional Multichannel System

Non Overlapping Adjacent Channels.

Channels separated by More

Than Their Two Sided bandwidth

OFDM Multichannel System

50% Overlap of Adjacent Channels

Available bandwidth is Used Twice

Channels separated by Half

Their Two Sided bandwidth
Standard Digital
Communication System
Bandlimited Channel
Nyquist Spectrum Nyquist Spectrum
With Cosine Taper

Infinite Duration Nyquist Pulse Finite Duration Nyquist Pulse

Translation in Time Domain
Phase Slope in Frequency Domain

Translated Signals Are Orthogonal When Peak is Translated to Zero Crossings of Original
Channel Distortion
Modifies Received Wave Shape
Matched Filter Output For SQRT Nyquist Pulse

1
0.8

0.6

0.4

0.2

0
-0.2
-6 -4 -2 0 2 4 6

Matched Filter Output For Channel Distorted SQRT Nyquist Pulse

1
0.8

0.6
0.4

0.2

-0.2
-6 -4 -2 0 2 4 6
Inter Symbol Interference (ISI)
Due to Channel Distorted Signal Eye Diagram No Channel Distortion
2

-1

-2
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1

Eye Diagram With Channel Distortion

2

-1

-2
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
Steady State Response of a Filter
to a Sine Wave is a Sine Wave
Rectangle Pulse: DC Centered Spectrum
with Equally Spaced Zeros

TP
sin(2 f )
H0 (f )  ATP 2
TP
(2 f )
2
Shift Spectrum with Linear Phase on DC
Pulse: Move Spectrum to First Spectral Zero

1 TP
sin( 2 ( f  k) )
Tp 2
H k ( f )  ATP
1 T
(2 ( f  k ) P )
Tp 2
Real Part of Complex Exponential Time Series:
Integer Number of Cycles per Interval
1 1 1 1

0 0 0 0

-1 -1 -1 -1
0 0.5 1 0 0.5 1 0 0.5 1 0 0.5 1
1 1 1 1

0 0 0 0

-1 -1 -1 -1
0 0.5 1 0 0.5 1 0 0.5 1 0 0.5 1
1 1 1 1

0 0 0 0

-1 -1 -1 -1
0 0.5 1 0 0.5 1 0 0.5 1 0 0.5 1
1 1 1 1

0 0 0 0

-1 -1 -1 -1
0 0.5 1 0 0.5 1 0 0.5 1 0 0.5 1
Imaginary Part of Complex Exponential Time Series:
Integer Number of Cycles per Interval
1 1 1 1

0 0 0 0

-1 -1 -1 -1
0 0.5 1 0 0.5 1 0 0.5 1 0 0.5 1
1 1 1 1

0 0 0 0

-1 -1 -1 -1
0 0.5 1 0 0.5 1 0 0.5 1 0 0.5 1
1 1 1 1

0 0 0 0

-1 -1 -1 -1
0 0.5 1 0 0.5 1 0 0.5 1 0 0.5 1
1 1 1 1

0 0 0 0

-1 -1 -1 -1
0 0.5 1 0 0.5 1 0 0.5 1 0 0.5 1
Spectra Of Complex Exponential Time Series:
Integer Number of Cycles per Interval
1 1 1 1

0.5 0.5 0.5 0.5

0 0 0 0

0 0.5 0 0.5 0 0.5 0 0.5

1 1 1 1

0.5 0.5 0.5 0.5

0 0 0 0

0 0.5 0 0.5 0 0.5 0 0.5

1 1 1 1

0.5 0.5 0.5 0.5

0 0 0 0

0 0.5 0 0.5 0 0.5 0 0.5

1 1 1 1

0.5 0.5 0.5 0.5

0 0 0 0

0 0.5 0 0.5 0 0.5 0 0.5

Continuous Time: Orthogonal Time Signal Set

k   ,  2,  1, 0 , 1, 2,  ,
 k (t ) :
0t T

2 k   ,  2,  1, 0 , 1, 2,  ,
 k (t )  exp( j   k t ) :
T 0t T

T
0 if n  m
 (t )  (t ) dt 


T if n  m
n m
0
Discrete Time: Orthogonal Time Signal Set
k  ...,  2,  1, 0 , 1, 2, ....,
 k (n ) :
0  n  N 1

2 k  0 , 1, 2, ...., N  1
 k (n )  exp( j   k nT ) :
NT 0  nT  NT
2 k  0 , 1, 2, ...., N  1
 exp( j   k n) :
N 0nN

N 1
0 if n  m
 n (n )  m (n )
n 0

N if n  m

NOTE :  k (n )   N  k (n )   k (n )   N k (n )
OFDM Modulator
OFDM Demodulator
OFDM is a Block Process
Adjacent Symbol Interference (ASI)
Symbol Smearing Due to Channel
Guard Interval Inserted Between Adjacent
Symbols to Suppress ASI
Cyclic Prefix Inserted in Guard Interval to
Suppress Adjacent Channel Interference (ACI)
Data Length Defines Sinc Width:
Spectral Spacing Matches Width
Extended Data Length Reduces Sinc
Width: Spectral Spacing Preserved
OFDM Symbol: Time and Spectra
Channel Input and Output
Real Part of Time Series, Input to Channel Spectrum
0.6 10

5
0.4
0
0.2
-5

0 -10

-15
-0.2
-20
-0.4
-25

-30
20 40 60 80 100 120 140 160 180 -0.5 0 0.5

Real Part of Time Series, Output of Channel Spectrum

0.6 10

5
0.4
0
0.2
-5

0 -10

-15
-0.2
-20
-0.4
-25

-30
20 40 60 80 100 120 140 160 180 -0.5 0 0.5
OFDM Spectra
Without and with Cyclic Prefix
OFDM spectral lines With Channel Without Cyclic Prefix
1.5

0.5

-0.5

-1

-1.5
-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5

OFDM spectral lines With Channel With Cyclic Prefix

1.5

0.5

-0.5

-1

-1.5
-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5
 Overlaid Constellations , All Frequencies,
Without and With Cyclic Prefix
OFDM Constellations With Channel Without Cyclic Prefix OFDM Constellations With Channel With Cyclic Prefix
1.5 1.5

1 1

0.5 0.5

0 0

-0.5 -0.5

-1 -1

-1.5
-1.5 -1 -0.5 0 0.5 1 1.5 -1.5
-1.5 -1 -0.5 0 0.5 1 1.5
Constellations: Different OFDM Bins
Without Cyclic Prefix

1 1 1

0 0 0

-1 -1 -1

-1 0 1 -1 0 1 -1 0 1

1 1 1

0 0 0

-1 -1 -1

-1 0 1 -1 0 1 -1 0 1
Constellations: Different OFDM Bins
With Cyclic Prefix

1 1 1

0 0 0

-1 -1 -1

-1 0 1 -1 0 1 -1 0 1

1 1 1

0 0 0

-1 -1 -1

-1 0 1 -1 0 1 -1 0 1
Channel Estimate with Pilots
Channel, Bandwidth, and Samples

0.5

0
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1

1
Zero Packed
Spectral Samples
0.5 and
Extended Reflection
0
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1

1
Interpolated
Spectral Points
0.5

0
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
0.01
Magnitude of Interpolation Error For In-Band Frequencies

0.005

0
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
Normalized Frequency
DFT (FFT) as Signal Generator
for Complex Sinusoids
DFT (FFT) As Signal Analyzer
for Complex Sinusoids

N 1 2
j
H (k )   h(n) e
nk
N
: k  0, 1, 2, ... , N  1
n 0
Radix-2 FFT Flow Diagrams
Input Vector FFT Mapped to Output Time Series,
Up-Sampled, Converted Via DAC to Waveform,
and I-Q Up-Converted
The FFT as Signal Generator
and Interpolator
OFDM Modulation With IFFT
and Interpolator
OFDM Demodulation With FFT
OFDM Transceiver
Time and Spectra of Sparse OFDM Symbol
Real Part OFDM Time Series
1

0.5
Amplitude

-0.5

-1
0 10 20 30 40 50 60 70 80 90 100
Normalized Time

Spectrum
10

0
Log Magnitude (dB)

-10

-20

-30

-40

-50

-60
-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5
Normalized Frequency
Time and Spectra With Frequency Offset = 0.1 Bin
Real Part OFDM Time Series with Offset Frequency = 0.1 Bin Width
1

0.5
Amplitude

-0.5

-1
0 10 20 30 40 50 60 70 80 90 100
Normalized Time

Spectrum With Frequency Offset = 0.1 Bin Width

10

-10
Log Magnitude (dB)

-20

-30

-40

-50

-60
-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5
Normalized Frequency
Time and Spectra With Sample Clock Offset = 1.02 fs
Real Part OFDM Time Series with Sampling Clock = 1.02 fs
1

0.5
Amplitude

-0.5

-1
0 10 20 30 40 50 60 70 80 90 100
Normalized Time

Spectrum With Sampling Clock = 1.02 fs

10

-10
Log Magnitude (dB)

-20

-30

-40

-50

-60
-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5
Normalized Frequency
Time and Spectra With Sample Clock Offset = 0.98 fs
Real Part OFDM Time Series with Sampling Clock = 0.98 fs
1

0.5
Amplitude

-0.5

-1
0 10 20 30 40 50 60 70 80 90 100
Normalized Time

Spectrum With Sampling Clock = 0.98 fs

10

-10
Log Magnitude (dB)

-20

-30

-40

-50

-60
-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5
Normalized Frequency
 Ideal I-Q Up and Down Conversion

c os(t) c os(t)
^
I(t) I(t)
Shape Match

CHANNEL

Shape n(t) Match

Q(t) ^
Q(t)
-sin(t) -sin(t)
Spectral and Time Description of
Real Sinusoids
Complex Sinusoids-I
Complex Sinusoids-II
Complex Baseband and Complex
Band-Centered Spectra
Complex Baseband and Real
Band-Centered Spectra
Complex Down Conversion
Gain and Phase Imbalance in
I-Q Mixers
Spectral Image Due to Gain Imbalance
Spectral Image Due to Phase Imbalance
Line Spectral Images Due to I-Q Mismatch
Coupling Between Positive and Negative FFT Indices Due to
I-Q Imbalance and First Order Correction Mechanism

    
 (1  j ) (   j )
 G(k )  2 2 2  H (k ) 
G( k )     

 
  (   j ) (1  j )   H ( k ) 
 2 2 2 

    
 (1  j ) (   j )
 H ( k )  2 2 2  G(k ) 
H ( k )      

 
  (   j ) (1  j )   G ( k ) 
 2 2 2 
Test Bench: Demonstration of Receiver I-Q
Imbalances, Carrier Offset, and Timing Offset
Carrier Offset: 4% of FFT Bin Width
Timing Offset: 10% of Sampling Time Period
Timing Clock Offset: 5% of Sampling Time Period per Frame
Gain Imbalance: 10% Error
Phase Imbalance: 0.1 Radian Error
I-Q Mixer Imbalance; 20% Gain, 0.2 Radians
Differential Delay to I/Q Mixers,
10% of Sample Interval
Periodic Time Segments in OFDM Frame
Obtained by Zero Packing Spectrum
Probe Mismatch During Short Repeated Preamble
Power Amplifier Non-Linearity
Nonlinear Transfer Function of Amplifier Input and Output of Non-Linear Amplifier
4 2

3.5 1.5

3 1

2.5 0.5

2 0

1.5 -0.5
1-dB Compression Point
1 -1

0.5 -1.5

0 -2
0 1 2 3 4 0 2 4 6 8 10

Spectrum of Two Input Sinusoids Spectrum of Two Output Sinusoids

10 10

0 0

-10 -10

-20 -20

-30 -30

-40 -40

-50 -50

-60 -60
-0.5 0 0.5 -0.5 0 0.5
Normalized Frequency Normalized Frequency
16-QAM Input and Output Envelopes.
Saturation and 1-dB Compression Circles
Saturation at 2-Times RMS Signal Level

Saturation
Envelope at Input to Amplifier Envelope at Output of Amplifier Saturation
2 2

1.5 1.5

1 1

0.5 0.5

0 0

-0.5 -0.5

-1 -1

-1.5 -1.5

-2 -2
-21-dB -1 0 1 2 -21-dB -1 0 1 2
Compression Compression
Limiting Amplifier Effect on Received QAM Constellation

Matched Filter Applied to Input of Amplifier Matched Filter Applied to Output of Amplifier

1 1

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0 0

-0.2 -0.2

-0.4 -0.4

-0.6 -0.6

-0.8 -0.8

-1 -1

-1 -0.5 0 0.5 1 -1 -0.5 0 0.5 1

Limiting Amplifier Effect on Signal Spectra
Spectrum at Input to Amplifier
10

-10
Log Magnitude (dB)

-20

-30

-40

-50

-60

-4 -3 -2 -1 0 1 2 3 4
Normalized Frequency (f/fsym)

Spectrum at Output of Amplifier

10

-10
Log Magnitude (dB)

-20
SPECTRAL REGROWTH
-30

-40

-50

-60

-4 -3 -2 -1 0 1 2 3 4
Normalized Frequency (f/fsym)
16-QAM (=0.2) Envelope Statistics
16-QAM Histogram at Amplifier Input 0
Probabilty of Level Crossing
10

0.016

0.014

0.012
-1
10
0.01

0.008

0.006
-2
0.004 10
0.002

0
0 0.5 1 1.5 2 2.5
Normalized Amplitude (x/  x )
-3
10

16-QAM Histogram at Amplifier Output

0.016 std dev =1.03 clip level

-4
0.014 10

0.012

0.01

0.008 -5
10
0.006

0.004

0.002
-6
0 10
0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5 3
Normalized Amplitude (x/  x ) Normalized Amplitude (x/  x )
OFDM Input and Output Envelopes:
Saturation and 1-dB Compression Circles
Saturation at 2-Times RMS Signal Level

Envelope at Input to Amplifier Envelope at Output of Amplifier

3 3
Saturation Saturation

2 2

1 1

0 0

-1 -1

-2 -2

1-dB 1-dB
-3 -3
Compression Compression
-3 -2 -1 0 1 2 3 -3 -2 -1 0 1 2 3
Limiting Amplifier Effect on OFDM Constellation

OFDM Constellation at Input to Amplifier OFDM Constellation at Output of Amplifier

1 1

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0 0

-0.2 -0.2

-0.4 -0.4

-0.6 -0.6

-0.8 -0.8

-1 -1

-1 -0.5 0 0.5 1 -1 -0.5 0 0.5 1

OFDM Envelope Statistics
OFDM Histogram at Amplifier Input 0
Probabilty of Level Crossing
0.015 10

0.01
-1
10

0.005
-2
10

0
0 1 2 3 4
Normalized Amplitude (x/  x )
-3
10
OFDM Histogram at Amplifier Output
0.015
std dev =1.06 clip level
-4
10

0.01

-5
10
0.005

-6
0 10
0 1 2 3 4 0 1 2 3 4
Normalized Amplitude (x/  x ) Normalized Amplitude (x/  x )
OFDM Envelope Statistics with
Selected Alternate Mapping
OFDM Histogram: One FFT 0
Probabilty of Level Crossing
0.02 10

0.015

0.01
-1
10
0.005

0
0 1 2 3 4
-2
OFDM Histogram: Two FFTs 10
0.02

0.015

Two FFTs One FFT

-3
0.01 10

0.005

0 Four FFTs
0 1 2 3 4 -4
10
OFDM Histogram: Four FFTs
0.02

0.015
-5
10
0.01

0.005

-6
0 10
0 1 2 3 4 0 1 2 3 4
Normalized Amplitude (x/  x ) Normalized Amplitude (x/  x )
Clipping
Smart Clipping
Reserve Frequency Bins Form Clipping
Pulses
Selecting Reserve Frequency Bins
Spectrum 11-Adjacent Frequencies Time Series for 11-Adjacent Frequencies
1 1

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0 0
-60 -40 -20 0 20 40 60 -0.5 0 0.5

Spectrum 11-Equally Spaced Frequencies Time Series for 11-Equally Spaced Frequencies
1 1

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0 0
-60 -40 -20 0 20 40 60 -0.5 0 0.5

Spectrum 11-Randomly Spaced Frequencies Time Series for 11-Randomly Spaced Frequencies
1 1

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0 0
-60 -40 -20 0 20 40 60 -0.5 0 0.5
Reserve Bin Canceller Clipping at 2.5  (8 dB)
Peak envelope, input to PAR control Peak envelope, output of first pass PAR control
5 5
data
clip level
4 4 data std dev
average peak

3 3

2 2

1 1

0 0
0 50 100 150 200 250 0 50 100 150 200 250

Peak envelope, output of second pass PAR control Peak envelope, output of third pass PAR control
5 5
data data
clip level clip level
4 data std dev 4 data std dev
average peak average peak

3 3

2 2

1 1

0 0
0 50 100 150 200 250 0 50 100 150 200 250
Statistics for Clip at 2.5 (8 dB)
input histogram 0
prob of level crossing
0.025 10
input
pass-1
0.02 pass-2
pass-3

-1
0.015 10

0.01

average =-0.648 dB
-2
0.005 10

0
0 1 2 3 4

-3
10

output histogram
0.025

std dev =0.928 -4

0.02 10

0.015

-5
0.01 10

clip level
0.005

-6
0 10
0 1 2 3 4 -5 0 5 10
PAR (dB)
Reserve Bin Canceller Clipping at 2.2  (6.9 dB)
Peak envelope, input to PAR control Peak envelope, output of first pass PAR control
5 5
data
clip level
4 4 data std dev
average peak

3 3

2 2

1 1

0 0
0 50 100 150 200 250 0 50 100 150 200 250

Peak envelope, output of second pass PAR control Peak envelope, output of third pass PAR control
5 5
data data
clip level clip level
4 data std dev 4 data std dev
average peak average peak

3 3

2 2

1 1

0 0
0 50 100 150 200 250 0 50 100 150 200 250
Statistics for Clip at 2.2 (6.9 dB)
input histogram 0
prob of level crossing
0.025 10
input
pass-1
0.02 pass-2
pass-3

-1
0.015 10

0.01

average =-0.653 dB
-2
0.005 10

0
0 1 2 3 4

-3
10

output histogram
0.025

std dev =0.928

-4
0.02 10

0.015

-5
0.01 10

clip level
0.005

-6
0 10
0 1 2 3 4 -5 0 5 10
PAR (dB)
Reserve Bin Canceller Clipping at 2.0  (6 dB)
Peak envelope, input to PAR control Peak envelope, output of first pass PAR control
5 5
data
clip level
4 4 data std dev
average peak

3 3

2 2

1 1

0 0
0 50 100 150 200 250 0 50 100 150 200 250

Peak envelope, output of second pass PAR control Peak envelope, output of third pass PAR control
5 5
data data
clip level clip level
4 data std dev 4 data std dev
average peak average peak

3 3

2 2

1 1

0 0
0 50 100 150 200 250 0 50 100 150 200 250
Statistics for Clip at 2.0 (6 dB)
input histogram 0
prob of level crossing
0.025 10
input
pass-1
0.02 pass-2
pass-3

-1
0.015 10

0.01

average =-0.659 dB
-2
0.005 10

0
0 1 2 3 4

-3
10

output histogram
0.025

std dev =0.927 -4

0.02 10

0.015

-5
0.01 10

clip level
0.005

-6
0 10
0 1 2 3 4 -5 0 5 10
PAR (dB)
OFDM 802.11a
Time-Frequency Profile of 802.11a Tones
Pilot Tones Shown in Yellow
Preamble and Pilot Structure

 Short Symbols
 Start of Frame Detection
 Signal Strength Indication
 Frequency Offset Resolution
 Long Symbols
 Channel Estimate
 Fine Time Resolution
 Distributed Pilots
 Carrier Tracking
 Sample Clock Tracking
Preamble Time Structure
Magnitude IEEE 802.11a Preamble
0.8

0.6

0.4

0.2

0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

Real Part
1

0.5

-0.5

-1
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

Imaginary Part
1

0.5

-0.5
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Detecting Frame Start with
Repeated Short Symbols
Signals in Preamble Detector
Envelope of Input Signal
0.8

0.6

0.4

0.2

0
0 0.5 1 1.5 2 2.5 3 3.5 4
Delayed Envelope of Input Signal
0.8

0.6

0.4

0.2

0
0 0.5 1 1.5 2 2.5 3 3.5 4
cross and auto correlations of Input Signal
0.8

0.6
Cross Correlation
0.4

0.2

0
0 0.5 1 1.5 2 2.5 3 3.5 4
Ratio of Cross to Auto Correlation
1

Detection
0.5
Threshold

0
0 0.5 1 1.5 2 2.5 3 3.5 4
Detail of Signal in Preamble Detector
Envelope of Input Signal
0.8

0.6

0.4

0.2

0
0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6
Delayed Envelope of Input Signal
0.8

0.6

0.4

0.2

0
0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6
cross and auto correlations of Input Signal
0.8

0.6
Auto Correlation
0.4

0.2
Cross Correlation
0
0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6
Ratio of Cross to Auto Correlation
1

Detection
0.5
Threshold

0
0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6
Maximum Likelihood Estimator for
Frequency Offset
Frequency and Signal Strength Estimates
Estimate of Frequency Offset
Frequency Offset in FFT Bins 10

Known Offset
3.3 FFT Bins
0

-5

-10
0 0.5 1 1.5 2 2.5 3 3.5 4
Sample Time

Estimate of Signal Strength

0.1

0.08
Mag Square

0.06

0.04

0.02

0
0 0.5 1 1.5 2 2.5 3 3.5 4
Sample Time
Cross Correlation of Long Preamble
Cross Correlation of Input Signal With Long Preamble Section
14

12

10
Amplitude

0
0 1 2 3 4 5 6 7
Time Samples

Zoom to First Correlation Peak Zoom to Second Correlation Peak

15 15

10 10
Amplitude

Amplitude

5 5
Expected Peak Position Expected Peak Position

0 0
4.3 4.4 4.5 4.6 4.7 4.8 5.3 5.4 5.5 5.6 5.7 5.8
Time Samples Time Samples
Clipped Cross Correlation of Long Preamble
Clipped Cross Correlation of Input Signal With Long Preamble Section
50

40

Clipped Correlator
30
Replica Signal Clipped Version of Template Signal
Amplitude

Sign[Real(Template)]+j*sign[Imag(Template)]

20

10

0
0 100 200 300 400 500 600 700 800 900
Time Samples

Zoom to First Correlation Peak Zoom to Second Correlation Peak

50 50

40 40

30 30
Amplitude

Amplitude

20 20

10 10

0 0
4.3 4.4 4.5 4.6 4.7 4.8 5.3 5.4 5.5 5.6 5.7 5.8
Time Samples Time Samples
Channel Probe With Long Preamble
Channel Probe, Long Segment of Preamble Response of Channel Probe, One Look
7 7

6 6

5 5
Amplitude

Amplitude
4 4

3 3

2 2

1 1

0 0
-0.5 0 0.5 -0.5 0 0.5
Normalized Frequency Normalized Frequency

Response of Channel Probe, No Noise Response of Channel Probe, Average of Two Looks
7 7

6 6

5 5
Amplitude

Amplitude
4 4

3 3

2 2

1 1

0 0
-0.5 0 0.5 -0.5 0 0.5
Normalized Frequency Normalized Frequency
Constellation with Residual Carrier Offset

5000 Constellations, Zero Carrier Frequency Offset 400 pilot, Zero Carrier Frequency Offset
1.5 1.5

1 1

0.5 0.5

0 0

-0.5 -0.5

-1 -1

-1.5 -1.5
-1.5 -1 -0.5 0 0.5 1 1.5 -1.5 -1 -0.5 0 0.5 1 1.5

5000 Constellations, 5 ppm Carrier Frequency Offset 400 pilot, 5 ppm Carrier Frequency Offset
1.5 1.5

1 1

0.5 0.5

0 0

-0.5 -0.5

-1 -1

-1.5 -1.5
-1.5 -1 -0.5 0 0.5 1 1.5 -1.5 -1 -0.5 0 0.5 1 1.5
Frequency Domain Residual Carrier Offset
60 Frames Real Part FFT: Zero Carrier Frequency Offset
1.5

0.5
Amplitude

-0.5

-1

-1.5
-30 -20 -10 0 10 20 30
Frequency Index

60 Frames Real Part FFT: 5 ppm Carrier Frequency Offset

1.5

0.5
Amplitude

-0.5

-1

-1.5
-30 -20 -10 0 10 20 30
Frequency Index
Constellations with Sample Clock Offset
5000 Constellations, Zero Clock Frequency Offset 400 pilot, Zero Clock Frequency Offset
1.5 1.5

1 1

0.5 0.5

0 0

-0.5 -0.5

-1 -1

-1.5 -1.5
-1.5 -1 -0.5 0 0.5 1 1.5 -1.5 -1 -0.5 0 0.5 1 1.5

5000 Constellations, 300 ppm Clock Frequency Offset 400 pilot, 300 ppm Clock Frequency Offset
1.5 1.5

1 1

0.5 0.5

0 0

-0.5 -0.5

-1 -1

-1.5 -1.5
-1.5 -1 -0.5 0 0.5 1 1.5 -1.5 -1 -0.5 0 0.5 1 1.5
Frequency Domain With Sample Clock Offset
80 Frames Real Part FFT: Zero Clock Frequency Offset
1.5

0.5
Amplitude

-0.5

-1

-1.5
-30 -20 -10 0 10 20 30
Frequency Index

80 Frames Real Part FFT: 300 ppm Clock Frequency Offset

1.5

0.5
Amplitude

-0.5

-1

-1.5
-30 -20 -10 0 10 20 30
Frequency Index
Other Variants of OFDM
Amplitude and Phase Overlays

 Shaped OFDM
 OQAM OFDM
 Coded OFDM
 CI OFDM
Shape to Control Spectral Side Lobes
Overlapped OFDM Frames
Polyphase Filter For Shaped OFDM
Shaping and Matched Filter
Impulse Response of Shaped OFDM
Modulator and Demodulator
1
Impulse at Input to IFFT (DC-bin)
0.5

0
0 1 2 3 4 5 6 7 8
1
Impulse Response at Output of IFFT
0.5

0
0 1 2 3 4 5 6 7 8
1
Impulse Response at Output of Polyphase Shaping Filter
0.5

0
0 1 2 3 4 5 6 7 8
1.5
Impulse Response at Output of
1
Polyphase Matched Filter
0.5
0
0 1 2 3 4 5 6 7 8
1
Impulse Response at Output of FFT
0.5

0
0 1 2 3 4 5 6 7 8
Orthogonal: Adjacent Time Slots Non
Adjacent Frequency Bins
Impulse Response Shaping Filter Auto Correlation Response

1 1

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0 0

-0.2 -0.2
0 1 2 3 4 5 0 2 4 6 8 10

Spectrum: Shaping Filter Centered on IFFT Spectral Bins

0
Adjacent Spectral Bins Correlated
Alternate Spectral Bins Not Correlated
-20

-40

-60

-4 -3 -2 -1 0 1 2 3 4
Impulse Response Time-Frequency Profile

0.8

0.6

0.4

0.2

0
8

6 10
OFD
M Fram 5
4
eN
u mb 0
er
2 c y bin
-5 frequ en
0 -10
Orthogonality Between Real and Imaginary
Part of Shaped OFDM Frequency Bins
Even and Odd Symmetric Wave Shapes from
Adjacent Bins are Orthogonal in Shaped OFDM
Symmetry Considerations in Real and Imaginary
Components of Offset Shaped OFDM Frames
Offset OFDM
Compare Spectra
Spectrum of Standard OFDM With Cyclic Prefix

0
Log Magnitude (dB)

-20

-40

-60

-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5
Normalized Frequency

Spectrum of OFDM/OQAM Without Cyclic Prefix

0
Log Magnitude (dB)

-20

-40

-60

-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5
Normalized Frequency
OFDM and Shaped OFDM PAR
Histogram: Standard OFDM 0
Probabilty of Level Crossing
10

0.016

0.014

0.012

0.01 -1
10
0.008

0.006

0.004

0.002 Shaped OFDM/OQAM

-2
10
0
0 1 2 3 4
Normalized Amplitude (x/  x )

Histogram: Shaped OFDM/OQAM

Standard OFDM
-3
10
0.016

0.014

0.012

0.01
-4
10
0.008

0.006

0.004

0.002

0
0 1 2 3 4 0 1 2 3 4
Normalized Amplitude (x/  x ) Normalized Amplitude (x/  x )
Complementary Codes
Canceling Correlation Side Lobes
Inserting CC in OFDM
AN (n )  AN (n )  AN (n )  AN (n )  2N  (n )
AN ( )  BN ( )  2N (A Constant Power Level)
2 2

Since Sample Values AN ( n ) are equal to 1

Average Power in AN ( )
2
=N
2N
Thus Peak to Average Power Ratio  2
N

Now Reverse Domains

Use Complementary Code Sequence
as amplitude of Carriers in Frequency Domain

Then time series has

Peak Squared Magnitude = 2N
Average Magnitude = N
for Peak to Average Power Ratio = 2
CC and Digital Filters
Equivalent Phase Coding
PAR in CCK OFDM and Standard OFDM
Mag Square of CC Rate 1/2 OFDM
2.5

Peak = 2
2

1.5

Mean = 1
1

0.5

0
0 500 1000 1500 2000 2500 3000 3500

Mag Square of Coded Rate 1/2 OFDM

8
Peak = 6.26

2
Mean = 1

0
0 500 1000 1500 2000 2500 3000 3500
OFDM and CC-OFDM PAR
C-I OFDM
Carrier Interferometry
 OFDM with Phase Overlay
 In Conventional OFDM
Rectangle Envelope in Time
Dirichlet Kernel in Frequency
 In CI-OFDM
Rectangle Envelope in Frequency
Dirichlet Kernel in Time
 Sin(x)/x in Time Domain Without Excess Bandwidth,
No Square-Root Nyquist Shaping Filter
Frequency Domain Phase Slope in
Continuous and in Sampled Data Domains
Circularly Shifted Time Domain Dirichlet Kernels

1 1 1 1 1

0.5 0.5 0.5 0.5 0.5

0 0 0 0 0
exp(-j 2 k 0/N) exp(-j 2 k 1/N) exp(-j 2 k 2/N) exp(-j 2 k 3/N) exp(-j 2 k 4/N)
-0.5 -0.5 -0.5 -0.5 -0.5
-10 0 10 -10 0 10 -10 0 10 -10 0 10 -10 0 10

1 1 1 1 1

0.5 0.5 0.5 0.5 0.5

0 0 0 0 0

exp(-j 2 k 5/N) exp(-j 2 k 6/N) exp(-j 2 k 7/N) exp(-j 2 k 8/N) exp(-j 2 k 9/N)
-0.5 -0.5 -0.5 -0.5 -0.5
-10 0 10 -10 0 10 -10 0 10 -10 0 10 -10 0 10

1 1 1 1 1

0.5 0.5 0.5 0.5 0.5

0 0 0 0 0
exp(-j 2 k 10/N) exp(-j 2 k 11/N) exp(-j 2 k 12/N) exp(-j 2 k 13/N) exp(-j 2 k 14/N)
-0.5 -0.5 -0.5 -0.5 -0.5
-10 0 10 -10 0 10 -10 0 10 -10 0 10 -10 0 10
Linear Versus Circular Convolution
Fast Circular Convolution with the FFT
Single Symbol in CI-OFDM
M-Symbols in CI-OFDM
1-to-2 Interpolated Time Domain Data Points
CI-OFDM Real Time Series and 1-Modulation Sample

1
0.8
Amplitude

0.6

0.4

0.2

-0.2
-20 -15 -10 -5 0 5 10 15 20
Time

CI-OFDM Real Time Series and 4-Modulation Samples

1
0.8
Amplitude

0.6

0.4

0.2

-0.2
-20 -15 -10 -5 0 5 10 15 20
Time
CI-OFDM Data Frame
CI-OFDM Real Time Series and 32-Modulation Samples

1.5

0.5
Amplitude

-0.5
Cyclic Prefix
-1

-1.5

-2
-25 -20 -15 -10 -5 0 5 10 15 20
Time

Spectrum
10

-10
Log Magnitude (dB)

-20

-30

-40

-50

-60
-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5
Normalized Frequency
CI-OFDM Statistics
Histogram: Standard OFDM (QPSK) 0
Probabilty of Level Crossing
10

0.015

0.01 -1
10

0.005

-2
10
0
0 1 2 3 4
Normalized Amplitude (x/  x )

Histogram: CI-OFDM (QPSK)

-3
10

0.1

-4
10

0.05

0
0 1 2 3 4 0 1 2 3 4
Normalized Amplitude (x/  x ) Normalized Amplitude (x/  x )
CI-OFDM Statistics
Histogram: Standard OFDM 0
Probabilty of Level Crossing
10

0.015

0.01 -1
10

0.005

-2
10
0
0 1 2 3 4
Normalized Amplitude (x/  x )

Histogram: CI-OFDM (16-QAM)

0.08 -3
10

0.06

0.04 -4
10

0.02

0
0 1 2 3 4 0 1 2 3 4
Normalized Amplitude (x/  x ) Normalized Amplitude (x/  x )
That’s all Folks

Professor harris, may I be excused?

My brain is full.