0% found this document useful (0 votes)
58 views15 pages

Multipath Can Be Described in Two Domains: Time and Frequency

Multipath can be described in time and frequency domains. In the time domain, it is represented by the impulse response. In the frequency domain, it is represented by the frequency response. Orthogonal frequency division multiplexing (OFDM) modulation encodes data in the frequency domain, with each frequency representing a carrier. During transmission, an inverse fast Fourier transform converts the signal to the time domain as a sum of sinusoids. At the receiver, a fast Fourier transform decodes each frequency bin independently. Loss of orthogonality due to a frequency offset between the transmitter and receiver causes interference between carriers. The interference increases quadratically with the offset value. A small offset, such as 0.001-
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
58 views15 pages

Multipath Can Be Described in Two Domains: Time and Frequency

Multipath can be described in time and frequency domains. In the time domain, it is represented by the impulse response. In the frequency domain, it is represented by the frequency response. Orthogonal frequency division multiplexing (OFDM) modulation encodes data in the frequency domain, with each frequency representing a carrier. During transmission, an inverse fast Fourier transform converts the signal to the time domain as a sum of sinusoids. At the receiver, a fast Fourier transform decodes each frequency bin independently. Loss of orthogonality due to a frequency offset between the transmitter and receiver causes interference between carriers. The interference increases quadratically with the offset value. A small offset, such as 0.001-
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 15

Multipath can be described in two domains: time and frequency

Time domain: Impulse response

time time
time

Impulse response

Frequency domain: Frequency response

time time

time f time

Sinusoidal signal as input Frequency response Sinusoidal signal as output


Introduction to OFDM modulation

N carriers
Symbol: 2 periods of f0

Transmit
f
+ f
Symbol: 4 periods of f0

B
Symbol: 8 periods of f0
Channel frequency
Data coded in frequency domain response
Transformation to time domain:
each frequency is a sine wave
in time, all added up.

N carriers Decode each frequency


bin separately
Receive
time f

Time-domain signal Frequency-domain signal


Introduction to OFDM (Orthogonal frequency division multiplex)
Data
Time-frequency grid
N carriers

Frequency
B
Carrier
f0

One OFDM symbol


Features T=1/f0
Time
– No intercarrier guard bands
– Controlled overlapping of bands Intercarrier Separation =
Any integer Multiple of 1/(symbol duration)
– Maximum spectral efficiency (Nyquist rate)
– Easy implementation using IFFTs
– Very sensitive to time-freq. synchronization

Modulation technique
A user utilizes all carriers simultaneously to transmit its data as coded quantity at each
frequency carrier, which can be quadrature-amplitude modulated (QAM).
OFDM Modulation and Demodulation using FFTs

b0 d0
b1 IFFT d1
P/S
b2 d0, d1, d2, …., dN-1
(Inverse fast d2
. (Parallel to
Fourier transform) d3
. serial converter)
. Transmit time-domain
. . samples of one symbol
f
. .
bN-1 .
time dN-1
Data coded in Data in time domain:
frequency domain: one symbol at a time
one symbol at a time

S/P d0’ b0’ Decode each


d0’, d1’, d2’, …., dN-1’ d1’ FFT b1’ frequency bin
(Serial to d2’ (Fast Fourier b2’ independently
parallel converter) d3’ transform) .
. .
Receive time-domain
. .
samples of one symbol f
. .
. bN-1’
time dN-1’
Loss of orthogonality (by frequency offset)
Transmission pulses ψ k (t) = exp( jk2π t / T ) y ψ k +m ( t) = exp ( j2 π (k + m)t / T )

Reception pulse with offset δ ψ δk +m (t) = exp (j2 π ( k + m + δ ) / T ) con δ ≤ 1 / 2

T (1 − exp(− j2πδ ))
exp( jk 2π t / T ) exp(− j(k + m + δ )2πt / T )dt =
T
Interference between
channels k and k+m
I m (δ ) = ∫
0 j2π (m + δ )

N −1
T sin πδ 1 23
Im (δ ) =
Summing up
∀m
∑ I 2m (δ ) ≈ (Tδ )2 ∑ 2 ≈( δ)
T 2 for N >> 1 (N > 5 Is enough )
π m+δ m m =1 m 14

Total ICI due to loss of orthogonality


Loss for 8 carriers
0
-10
-10 m=1 -15 δ =0.05
-20
Interference in dB

-20 m=3 -25 δ =0.02


δ =0.01
m=5 ICI in dB
-30 m=7 -30

-40
-35 δ =0.005
-40
Practical limit
-50 -45 δ =0.002
-50 δ =0.001
-60 -55
-70 -60
0 2 4 6 8 10 12 14 16
-0.4 -0.3 -0.2 -0.1 0.1 0.2 0.3 0.4
Carrier position within the band (N=16)
Frequency offset
δ assumed r.v.
Gaussian σ=δ
Loss of orthogonality (time)

− T / 2+ τ T/ 2 2 consecutive
Let us assume Xi = c0 ∫ ψ k (t)ψ *l (t − τ )dt + c 1 ∫−T / 2 +τ ψ k (t)ψ l* (t − τ )dt
a misadjustment τ −T /2 symbols

 τ
senmπ τ
2T T , c ≠c
Then Xi =  mπ
0 1
Or approximately, Xi 2mπ T τ independent
≈ =2
 when τ<<T T mπ T on m
if m=k-l  0, c0 = c1

 τ
In average, the interfering  Xi 2  τ 1
2
1  τ
2 ICI ≈ 20 log  2  , τ << T
E 2 = + =
 4  T  2 0 2 2 T 
T
power in any carrier is
T  Per carrier

Loss for 16 carriers ICI due to loss of orthogonaliy


0 -45
-5
m=1 -40
-10
Interference in dB

-15 -35
assumed a uniform r.v.
-20
ICI in dB

m=5 -30
-25
-30 m=10 -25 N=8 Max. practical limit
-35 -20
-40 N=64
-15
-45
-50 -10
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08

Relative misadjustment τ Standard deviation of the relative misadjustment


Zone of interest
Including a cyclic prefix to each OFDM symbol

To combat the multipath: including time guards between the symbols

copy

CP
τ T
Tc

Without the Cyclic Prefix Including the Cyclic Prefix


Symbol: 8 periods of fi
CP Symbol: 8 periods of fi
Passing the channel h(n)

Passing the channel h(n)


Ψi(t)
Ψi(t)

Channel: h(n)=(1 ) –n / n n=0 ,…,2 3

≠Ψi(t)

Initial transient The inclusion of a CP Final transient


remains within maintains the orthogonality remains within
Initial transient Loss of orthogonality Decaying transient
the CP the CP

Ψj(t) Ψj(t)

Symbol: 4 periods of fi Symbol: 4 periods of fi

CP functions:
– It acomodates the decaying transient of the previous symbol
– It avoids the initial transient reachs the current symbol
..

Symplified scheme of an OFDM transceiver

Transmitter

Cyclic
prefix (CP)

CODER
S

IFFT
BITS S RF
P DAC
.
P

f0

Receiver

DECOD.
P S BITS

FFT
RF ADC Filter
S P

f0 PLL,
symbol
Channel estimation
timing
frequency offset
Windowing of the OFDM symbol
Total band used by OFDM: it depends on the number or carriers

ACI

ACI

Wide separation Narrow


separation

Maintainig a fix bandwidth, if N increases Adjacent channel interference decreases

BUT
It is interesting to have few carriers as well:
• To introduce short delay in data gathering and signal processing (FFTs)
• To have a bigger intercarrier separation --> It reduces the relative frequency offset

Compromise
Need to shape the OFDM symbols
OFDM modulators with symbol shaping

p(t) ejωnt Equivalent ejωnt


architectures p(t)
an an

……… Σan p(t) ejωnt ……… p(t) Σanejωnt

ak ak

p(t) ejωkt ejωkt


Implemented with FFT

After the synchronous reception


p(t) even
T/2 T/2 0 , k ≠ n even
I=∫ dt =2 ∫ p(t)cos[2π (n − k)t / T]dt = 
j2 π (k− n)t / T p(t)
p(t)e
−T / 2 0 1 , k = n 0
T/2

cos
The simplest way to maintain symmetry within -T/2<t<T/2 is p(t)=k odd

❏ Symbol shaping has to be carried out as part PC+T


of the symbol duration + CP
❏ The total ACI can be condiderably reduced
Robustness against the channel and ACI improvement

……
Virtual OFDM symbols within the slot

• With guards (Cyclic prefixes), the channel’s time dispersion is avoided


L=N+CP

PC PC PC …… PC PC

OFDM symbols with time guards (CPs)

• With smooth transitions between symbols, the adjacent channel interference is minimized
PC PC PC …… PC

OFDM symbols with time guards and symbol shaping


802.11a Physical Layer Data Symbol Format

t1 t2 t3 t4 t5 t6 t7 t8 t9 t10 GI2 T1 T2 GI OFDM Symbol GI OFDM Symbol

Short training sequence: Long training sequence:


AGC and frequency offset Channel estimation

Training symbols: 4 us each Data rate for each 20 Mhz channel:


t: 0.8 us, 16 samples 20 Msamples per second.
GI2: 1.6 us, 32 samples 250 Ksymbols per second.
T: 3.2 us, 64 samples 48 data carriers per symbol.
1/2 or 3/4 convolutional code.
Data Symbols: 4 us each 1 bit/carrier (BPSK) to 6 bits/carrier (64 QAM).
GI: 0.8 us, 16 samples
OFDM Symbol: 3.2 us, 64 samples Overall:
Lowest: 48 * 1 * 1/2 * 250K = 6 Mbps.
* Only 52 of the 64 carriers are used. Highest: 48 * 6 * 3/4 * 250K = 54 Mbps.
* 4 of the 52 carriers are used for
pilot carriers (no data). Turbo mode supports 108 Mbps using 40 Mhz channel.
Robustness against errors: random noise and channel-selected errors

Random noise: primarily introduced by thermal and circuit noise.

Channel-selected errors: introduced by magnitude distortion in


channel frequency response.
Data bits
Time-frequency grid

Frequency
B
Bad carriers
f0

f Time
Frequency response T=1/f0

Errors are no longer random. Interleaving is often used to scramble


the data bits so that standard error correcting codes can be applied.
Spectrum Mask

Power Spectral Density

-20 dB

-28 dB

-40 dB

-30 -20 -11 -9 9 11 20 30


f carrier
Frequency (MHz)

• Requires extremely linear power amplifier design.


Adjacent Channel and Alternate Adjacent Channel Rejection
D a te M in im u m A d ja c e n t C h a n n e l A lte r n a te
ra te S e n s ib ility R e je c tio n C h a n n e l re je c tio n
6 M bps -8 2 d B m 16 dB 32 dB

12M bps -7 9 d B m 13 dB 29 dB

24M bps -7 4 d B m 8 dB 24 dB

36M bps -7 0 d B m 4 dB 20 dB

54M bps -6 5 d B m 0 dB 15 dB

32 dB

16 dB

Signal Frequency

• Requires joint design of the anti-aliasing filter and ADC.

You might also like