You are on page 1of 34

Wireless Information Transmission System Lab.

Introduction to Orthogonal Frequency


Division Multiplexing (OFDM) Technique

Institute of Communications Engineering National Sun Yat-sen University


Outline
OFDM Overview
OFDM System Model
Orthogonality
Multi-carrier Equivalent Implementation by Using IDFT
(IFFT)
Cyclic Prefix (CP)
Summary

2
OFDM Overview (1/3)
Single carrier (SC) vs. multi-carrier (MC)

Single carrier : data are Multi-carrier : data are shared


transmitted over only among several carriers and
one carrier simultaneously transmitted
Selective fading Flat fading per subcarrier

3
OFDM Overview (2/3)
OFDM
Orthogonal Frequency Division Multiplexing

The basic principle of OFDM is to split a high-rate data


stream into a number of lower rate streams that are
transmitted simultaneously over a number of sub-carriers.

It eliminates or alleviates the problem of multi-path


channel fading effect, low spectrum efficiency, and
frequency selective fading.

4
OFDM Overview (3/3)
OFDM modulation

Features
No intercarrier guard bands
Overlapping of bands
Spectral efficiency
Easy implementation by IFFTs
Very sensitive to synchronization

5
OFDM Transmitter

cos(2πf 0 t )

a(0)

b(0 )

• sin (2πf 0 t )
Data d (n ) D(t )
Input Encoder S/P • MUX Channel
fs =
1 {a(n ) + jb(n )} • cos(2πf N −1t )
Δt
1 a( N − 1)
f n = f 0 + nΔf , Δf =
NΔt
b( N − 1)

sin (2πf N −1t )

Figure 1

6
Multi-carrier Systems
Multi-carrier Transmission

8T
7
OFDM System Model (1/3)
An OFDM system transmitter shown in Figure 1.
The transmitted waveform D(t) can be expressed as
N −1
D (t ) = ∑ {a ( n ) cos( 2πf n t ) + b( n) sin( 2πf n t )} (1)
n =0

1
where f n = f 0 + nΔf and Δf =
N Δt

Using a two-dimensional digital modulation format, the data


symbols d(n) can be represented as a(n) + jb(n)
a(n) : in-phase component
b(n) : quadrature component

8
OFDM System Model (2/3)

NΔt
+1
a(0 ) cos(2πf 0 t )
t a(0)
T
b(0)

• sin (2πf 0 t )
Data d (n ) D(t )
Input Encoder S/P • MUX Channel
fs =
1 {a(n ) + jb(n )} • cos(2πf N −1t )
Δt
a( N − 1)
a(0 )
+1 b(N − 1)
Δt T
t sin (2πf N −1t )
−1
a(1) a (N − 1)

NΔt
9
OFDM System Model (3/3)
The serial data elements spaced by Δt are grouped and
used to modulate N carriers. Thus they are frequency
division multiplexed.

The signaling interval is then increased to NΔt , which


makes the system less susceptible to channel delay
spread impairments.
Small-scale fading
(Based on multipath time delay spread)

Flat Fading Frequency Selective Fading


1. BW of signal < BW of channel 1. BW of signal > BW of channel
2. Delay spread < Symbol preiod 2. Delay spread > Symbol preiod

10
Orthogonality (1/3)
Consider a set of transmitted carriers as follows:
⎛ n ⎞
j 2π ⎜ f 0 + ⎟t
ψ n (t ) = e ⎝ N Δt ⎠
for n = 0, 1, ..., N − 1 (2)

We now show that every two carriers are orthogonal to


each other. <proof is in next page>

In summary:
b ⎧(b − a ) for p = q
∫aψ p (t )ψ (t )dt = ⎨⎩ 0 for p ≠ q and (b − a ) = NΔt
*
q

11
Orthogonality (2/3) --- Proof
t
b b j 2π ( p − q )
∫ψ (t )ψ (t ) dt = ∫ e N Δt
dt
*
p q
a a
b a
j 2π ( p − q ) j 2π ( p − q )
e −e N Δt N Δt
=
j 2π ( p − q ) NΔt

j 2π ( p − q )
b
j 2π ( p − q )
1
( a −b ) ⎞
e ⎜1 − e N Δt N Δ t ⎟
⎜ ⎟
= ⎝ ⎠
j 2π ( p − q ) NΔt
= 0 , for p ≠ q and (b − a ) = NΔt

12
Orthogonality (3/3)

NΔt

13
Frequency Error Results in Inter-Carrier
Interference (ICI)

0.8

0.6

Inter-Carrier Interference
0.4

0.2

-0.2

-0.4
-3 -2 -1 0 1 2 3

14
Timing Synchronization Error Results in
Inter-Symbol Interference (ISI)

Not Orthogonal Any More.

15
Mathematical Expression of
OFDM Signal (1/2)
From above, we know that {ψ n (t )} is the orthogonal
signal set. An OFDM signal based on this orthogonal
signal set can be written as:
⎧ ∞ N −1 ⎫
x (t ) = Re ⎨ ∑ ∑ d k ,nψ n (t − kT ) ⎬ (3)
⎩k = −∞ n = 0 ⎭
where ψ n (t ) = e j 2πf nt for n = 0, 1, 2, ..., N − 1 0 ≤ t ≤ T
n
f n = f 0 + , T = N Δt
T
d k ,n = ak ,n + jbk ,n

16
Mathematical Expression of
OFDM Signal (2/2)
T : OFDM symbol duration
dk,,n : transmitted data on the n-th carrier of the k-th symbol

⎧ ∞ N −1 ⎫
x (t ) = Re ⎨ ∑ ∑ C k , nψ n (t − kT ) ⎬
⎩k = −∞ n = 0 ⎭
∞ N −1
= ∑ ∑ {a
k = −∞ n = 0
k ,n cos (2πf n (t − kT ) ) − bk ,n sin (2πf n (t − kT ) )} (4)

If there is only one OFDM symbol ( i.e. k = 0 ), it can be


simplified as:
N −1
x(t ) = ∑ { an cos(2πf nt ) − bn sin( 2πf nt ) } (5)
n=0

17
Multi-carrier Equivalent
Implementation by using IDFT (1/6)
According to the structure of Tx, it must use N
oscillators. That increases the hardware complexity.

The equivalent method is using IDFT (IFFT).

18
Multi-carrier Equivalent
Implementation by using IDFT (2/6)
In general, each carrier can be expressed as:

S c (t ) = Ac (t )e j (2πf c t +φc ( t ) ) (6)

We assume that there are N carriers in the OFDM signal.


Then the total complex signal Ss(t) can be represented by:
N −1
1
S s (t ) =
N
∑ n
A (
n =0
t ) e j ( 2πf n t +φn ( t ) )
(7)

where f n = f 0 + nΔf
and An (t ), φn (t ), f n are amplitude, phase, carrier frequency
of n - th carrier, respectively.
19
Multi-carrier Equivalent
Implementation by using IDFT (3/6)
Then we sample the signal at a sampling frequency 1 Δt ,
and An(t) and φn(t) becomes:
φ n (t ) = φ n (8)
An (t ) = An (9)
N −1
1
S s (kΔt ) =
N
∑Ae
n =0
n
j ( 2π ( f 0 + nΔf ) kΔt +φn )
(10)

Then the sampled signal can be expressed as:

( )⋅ e
N −1
1
S s (kΔt ) =
N

n =0
A en
( j 2πf 0 kΔt +φn ) j 2πnkΔfΔt
(11)

20
Multi-carrier Equivalent
Implementation by using IDFT (4/6)
The inverse discrete Fourier transform (IDFT) is defined
as the following:
1 N −1
f ( kΔt ) = ∑ F ( nΔf )e j 2πnk N (12)
N n=0

Comparing eq.(11) and eq.(12), the condition must be


satisfied in order to make eq.(11) an inverse Fourier
transform relationship:
1
Δf = (13)
NΔt

21
Multi-carrier Equivalent
Implementation by using IDFT (5/6)
If eq.(13) is satisfied,
An e j (2πf 0 kΔt +φn ) is the frequency domain signal
S s ( kΔt ) is the time domain signal
Δf is the sub-channel spacing
NΔt is the symbol duration in each sub-channel

This outcome is the same as the result obtained in the


system of Figure 1. Therefore IDFT can be used to
generate an OFDM transmission signal.

22
Multi-carrier Equivalent
Implementation by using IDFT (6/6)
cos(2πf 0 t )

a(0)

b(0)

• sin (2πf 0 t )
d (n ) D(t )
Input • Channel
fs =
1 {a(n ) + jb(n )}
Δt • cos(2πf N −1t )

1 a( N − 1)
f n = f 0 + nΔf , Δf =
N Δt
b( N − 1)

sin (2πf N −1t )


d (0)
d (1)
d (2)

• •
d (n ) D(t )
Input • • Channel
fs =
1 {a(n ) + jb(n )}
Δt • •
1 • •
f n = f 0 + nΔf , Δf = • •
NΔf
d (N − 1)

23
Cyclic Prefix (1/9)
In multipath channel, delayed replicas of previous OFDM signal
lead to ISI between successive OFDM signals.

Solution : Insert a guard interval between successive OFDM


signals.

24
Cyclic Prefix (2/9)
Guard interval leads to intercarrier interference (ICI) in OFDM
demodulation

In DFT interval, difference between two subcarriers does not


maintain integer number of cycles Î loss of orthogonality.
Delayed version of subcarrier 2 causes ICI in the process of
demodulating subcarrier 1.
25
Cyclic Prefix (3/9)
Cyclic prefix (CP) : A copy of the last part of OFDM signal is
attached to the front of itself.

d [0] D[0]
d [1] D[1]
d [2] D[2]
D[k ]
~

Input data d [n]








symbols •

[ ]



D N − Ng
• •
• •

D[N − 1]

d [N − 1]

26
Cyclic Prefix (4/9)
All delayed replicas of subcarriers always have an
integer number of cycles within DFT interval Î no ICI

27
Cyclic Prefix (5/9)
Linear convolution vs. circular convolution

28
Cyclic Prefix (6/9)
Channel effect with cyclic prefix

Signal after removed CP

29
Cyclic Prefix (7/9)
Spectrum of channel response h[n ] with length Lh (smaller
than N g )
H k = FFT {h[n]}

Received complete OFDM signal


r [n] = D[n] ⊗ h[n] , 0 ≤ n ≤ N + N g + Lh − 2
~ ~
Received useful part r [n]
r [n] = D[n] ⊗ N h[n]
where ⊗ N is N-point circular convolution (due to CP)
Received symbol at k-th subcarrier
Yk = FFT {r [n]} = FFT {D[n] ⊗ N h[n]} = X k H k
Y “Useful property for OFDM system to reduce
⇒ Xk = k
Hk complexity of channel equalization”

30
Cyclic Prefix (8/9) --- OFDM Receiver

• •
r (n ) Output data
r (n )
~
• •
Symbols
• •
• •
• •

31
Cyclic Prefix (9/9)
One of the most important reasons to do OFDM is the
efficient way it deals with multipath delay spread.
To eliminate inter-symbol interference (ISI) almost
completely, a guard time is introduced for each OFDM
symbol.
(The guard time is chosen larger than the delay spread)

32
Bandwidth Efficiency
In a classical parallel system, the channel is divided into
N non-overlapping sub-channels to avoid inter-carrier
interference (ICI).

The diagram for bandwidth efficiency of OFDM system


is shown below:

33
Summary
The advantage of the FFT-based OFDM system :
The use of IFFT/FFT can reduce the computation complexity.
The orthogonality between the adjacent sub-carriers will make
the use of transmission bandwidth more efficient.
The guard interval is used to resist the inter-symbol interference
(ISI).
The main advantage of the OFDM transmission technique is its
high performance even in frequency selective channels.
The drawbacks of the OFDM system :
It is highly vulnerable to synchronization errors.
Peak to Average Power Ratio (PAPR) problems.

34

You might also like