ECS455 - 4-6 - All
Uploaded by
Ammar IkramECS455 - 4-6 - All
Uploaded by
Ammar IkramECS455: Chapter 4
Multiple Access
Office Hours:
BKD 3601-7
Monday 9:20-10:20
Dr.Prapun Suksompong Wednesday 9:20-10:20
1 prapun.com/ecs455
2
ECS455: Chapter 4
Multiple Access
4.1 TDD and FDD
Office Hours:
BKD 3601-7
Monday 9:20-10:20
Dr.Prapun Suksompong Wednesday 9:20-10:20
3 prapun.com/ecs455
Duplexing
Allow the subscriber to send “simultaneously” information to the
base station while receiving information from the base station.
Talk and listen simultaneously.
Definitions:
Forward channel or downlink (DL) is used for communication
from the infrastructure to the users/stations
Reverse channel or uplink (UL) is used for communication from
users/stations back to the infrastructure.
Two techniques
1. Frequency division duplexing (FDD)
2. Time division duplexing (TDD)
4 [Rappaport, 2002, Ch 9]
Frequency Division Duplexing (FDD)
Provide two distinct bands of frequencies (simplex
channels) for every user.
The forward band provides traffic from the base station to
the mobile.
The reverse band provides traffic from the mobile to the
base station.
Any duplex channel actually consists of two simplex channels (a
forward and reverse).
Most commercial cellular systems are based on FDD.
5
FDD Examples
GSM
[Karim and Sarraf, 2002, Fig 5-1]
UMTS
[Karim and Sarraf, 2002, Fig 6-1]
6
Time Division Duplexing (TDD)
The UL and DL data are transmitted on the same carrier
frequency at different times. (Taking turns)
Use time instead of frequency to provide both forward and
reverse links.
Each duplex channel has both a forward time slot and a
reverse time slot.
If the time separation between the forward and reverse time
slot is small, then the transmission and reception of data
appears simultaneous to the users at both the subscriber unit
and on the base station side.
Used in Bluetooth and Mobile WiMAX
7
Problems of FDD
Each transceiver simultaneously transmits and receives radio
signals
The signals transmitted and received can vary by more than 100
dB.
The signals in each direction need to occupy bands that are
separated far apart (tens of MHz)
A device called a duplexer is required to filter out any
interference between the two bands.
8
[Tse and Viswanath, 2005, Ch 4, p 121]
Advantages of FDD
TDD frames need to incorporate guard periods equal to
the max round trip propagation delay to avoid
interference between uplink and downlink under worst-case
conditions.
There is a time latency created by TDD due to the fact that
communications is not full duplex in the truest sense.
This latency creates inherent sensitivities to propagation delays
of individual users.
9
Advantages of TDD
Duplexer is not required.
Enable adjustment of the downlink/uplink ratio to efficiently
support asymmetric DL/UL traffic.
With FDD, DL and UL always have fixed and generally, equal DL and
UL bandwidths.
Assure channel reciprocity for better support of link adaptation,
MIMO and other closed loop advanced antenna technologies.
Ability to implement in nonpaired spectrum
FDD requires a pair of channels
TDD only requires a single channel for both DL and UL providing
greater flexibility for adaptation to varied global spectrum
allocations.
10
ECS455: Chapter 4
Multiple Access
4.2 Introduction to Multiple Access
11
12
Multiple Access Techniques
Allow many mobile users to share simultaneously a finite
amount of radio spectrum.
For high quality communications, this must be done without
severe degradation in the performance of the system.
Important access techniques
1. Frequency division multiple access (FDMA)
2. Time division multiple access (TDMA)
3. Spread spectrum multiple access (SSMA)
Frequency Hopped Multiple Access (FHMA)
Code division multiple access (CDMA)
4. Space division multiple access (SDMA)
5. Random access
ALOHA
13
Chapter 4
Multiple Access
4.3 FDMA and TDMA
14
Frequency division multiple access
(FDMA)
The oldest multiple access scheme for wireless
communications.
Used exclusively for multiple access in 1G down to
individual resource units or physical channels.
Assign individual channels to individual users.
Different carrier frequency is assigned to each user so that the
resulting spectra do not overlap.
During the period of the call, no other user can share the same
channel.
Band-pass filtering (or heterodyning) enables separate
demodulation of each channel.
15
FDMA (2)
If an FDMA channel is not in use, then it sits idle and cannot
be used by other users to increase or share capacity.
It is essentially a wasted resource.
In FDD systems, the users are assigned a channel as a pair of
frequencies.
16 [Rappaport, 2002, Ch 9, p. 449]
Time division multiple access (TDMA)
Divide the radio spectrum into time slots.
In each slot only one user is allowed to either transmit or receive.
A channel may be thought of as a particular time slot that reoccurs
every frame, where N time slots comprise a frame.
Transmit data in a buffer-and-burst method
The transmission for any user is non-continuous.
Digital data and digital modulation must be used with TDMA.
This results in low battery consumption, since the subscriber
transmitter can be turned off when not in use (which is most of the
time).
An obvious choice in the 1980s for digital mobile
communications.
17
FDMA vs. TDMA
18
Tradeoffs
TDMA transmissions are slotted
Require the receivers to be synchronized for each data burst.
Guard times are necessary to separate users. This results in larger
overheads.
FDMA allows completely uncoordinated transmission in the
time domain
No time synchronization among users is required.
The complexity of FDMA mobile systems is lower when
compared to TDMA systems, though this is changing as digital
signal processing methods improve for TDMA.
Since FDMA is a continuous transmission scheme, fewer bits are
needed for overhead purposes (such as synchronization and
framing bits) as compared to TDMA.
FDMA needs to use costly bandpass filters.
For TDMA, no filters are required to separate individual physical
channels.
19
Guard Band vs. Guard Time
FDMA TDMA
20
Example: GSM
FDMA
GSM FDMA/TDMA
with one active time
slot
[Figure 7.2, Heine, 1998]
GSM utilizes a combination of Each narrowband channel has
FDMA and TDMA bandwidth 200 kHz.
Two-dimensional channel Time is divided into slots of length
21
structure T =577 s.
The FDMA/TDMA structure of GSM
In full-rate configuration, eight time slots (TSs) are mapped
on every frequency.
A BS with 6 carriers, as shown here, has 48 (8 times 6)
physical channels (in fullrate configuration).
22
[Figure 7.1, Heine, 1998]
Classifications of Medium Access
Control (MAC)
23
24
ECS455: Chapter 4
Multiple Access
4.4 DS/SS
Office Hours:
BKD 3601-7
Monday 9:20-10:20
Dr.Prapun Suksompong Wednesday 9:20-10:20
1 prapun.com/ecs455
Spread spectrum (SS)
Historically spread spectrum was developed for secure
communication and military uses.
Difficult to intercept for an unauthorized person.
Easily hidden. For an unauthorized person, it is difficult to even
detect their presence in many cases.
Resistant to jamming.
Provide a measure of immunity to distortion due to multipath
propagation.
In conjunction with a RAKE receiver, can provide coherent
combining of different multipath components.
Asynchronous multiple-access capability.
Wide bandwidth of spread spectrum signals is useful for location
and timing acquisition.
[Goldsmith, 2005, Ch 13]
2
Spread spectrum: Applications
First achieve widespread use in military applications due to
its inherent property of hiding the spread signal below the noise
floor during transmission,
its resistance to narrowband jamming and interference, and
its low probability of detection and interception.
The narrowband interference resistance has made spread
spectrum common in cordless phones.
The basis for both 2nd and 3rd generation cellular systems
as well as 2nd generation wireless LANs (WLAN).
The ISI rejection and bandwidth sharing capabilities of spread
spectrum are very desirable in these systems
[Goldsmith, 2005, Ch 13]
3
Spread spectrum conditions
Spread spectrum refers to any system that satisfies the following
conditions [Lathi, 1998, p 406 & Goldsmith, 2005, p. 378]:
1. The spread spectrum may be viewed as a kind of modulation
scheme in which the modulated (spread spectrum) signal
bandwidth is much greater than the message
(baseband) signal bandwidth.
2. The spectral spreading is performed by a code that is
independent of the message signal.
This same code is also used at the receiver to despread the received
signal in order to recover the message signal (from the spread
spectrum signal).
In secure communication, this code is known only to the person(s)
for whom the message is intended.
[R. Pickholtz, D. Schilling, L. Milstein, “Theory of Spread-Spectrum Communications - A Tutorial,” IEEE Trans. Commun., Vol. 30 , pp.
4
855-884, May 1982.]
Spread spectrum (2)
Increase the bandwidth of the message signal by a factor N, called
the processing gain (or bandwidth spreading factor).
In practice, N is on the order of 100-1000. [Goldsmith, 2005, p 379]
N = 128 for IS-95 [T&V]
Wasteful?
Although we use much higher BW for a spread spectrum signal,
Multiplexing: we can also multiplex large numbers of such signals
over the same band.
Multiple Access: many users can share the same spread spectrum
bandwidth without interfering with one another.
Achieved by assigning different code to each user.
Frequency bands can be reused without regard to the separation distance of
the users.
5
Spread Spectrum (3)
Two forms of spread spectrum (SS)
1. Frequency Hopping (FH)
Hop the modulated data signal over a wide BW by changing its
carrier frequency
BW is approximately equal to NB
N is the number of carrier frequencies
available for hopping
B is the bandwidth of the data signal.
The most celebrated invention of frequency hopping was that of
actress Hedy Lamarr
and composer George Antheil in 1942
2. Direct Sequence (DS)
6
7
For several years beginning in 1997, the boxes of
CorelDRAW's software suites were graced by a large Corel-
drawn image of Hedy Lamarr.
8
SSMA, CDMA, DS/SS
Single User Multi-access
SSMA
TDMA
FH/SS CDMA
FDMA
DS/SS
SDMA
9
Useful even for single user!
DS/SS System
𝑦 𝑡 =𝑚 𝑡 𝑐 𝑡 (Integrator)
y t c t m t c2 t m t
Code
generator Code 1
Synchronization/ Code
tracking generator
(Correlation)
Message signal
(data/information signal)
Pseudonoise (PN)
sequence. (Think of this as
a pseudorandom carrier).
Here, we refer to it as spreading code/sequence.
Tb
N
10 Tc
DS/SS System (Con’t)
Observe that…
To be able to perform the despreading operation, the
receiver must
know the code sequence c(t) used at the Tx to spread the
signal
synchronize the codes of the received signal and the locally
generated code.
The process of detection (despreading) is identical to the
process of spectral spreading.
Recall that for DSB-SC, we have a similar situation in that the
modulation and demodulation processes are identical (except
for the output filter).
11
Spread spectrum modem
[Viterbi, 1995, Fig. 1.2]
12
DS/SS: Spectral Spreading Signal c(t)
A pseudorandom signal
Appear to be unpredictable
Can be generated by deterministic means
(hence, pseudorandom)
The bit rate is chosen to be much higher then the bit rate of m(t).
The basic pulse in c(t) is called the chip.
The bit rate of c(t) is known as the chip rate.
The autocorrelation function of c(t) should be very narrow.
Small similarity with its delayed version
Remark: In multiuser (CDMA) setting, the cross-correlation
between any two codes c1(t) and c2(t) should also be very small
Negligible interference between various multiplexed signals.
13
Frequency-Domain Analysis
Shifting Properties: g t t0 e j 2 ft0 G f e j 2 f0t g t G f f0
1 1
Modulation: m t cos 2 f ct M f fc M f fc
2 2
14
DS/SS: Secure Communication
Secure communication
Signal can be detected only by authorized person(s) who know the
pseudorandom code used at the transmitter.
Signal spectrum is spread over a very wide band, the signal PSD is
very small, which makes it easier to hide the signal within the noise
floor
15
DS/SS: Jamming Resistance
y t i t c t m t c t i t c t m t i t c t
2
Jamming Resistance / Narrowband Interference rejection
The decoder despreads the signal y(t) to yield m(t).
The jamming signal i(t) is spread to yield i(t)c(t).
Using a LPF, can recover m(t) with only a small fraction of the power
from i(t).
Caution: Channel noise will not spread.
16
DS/SS: Multipath Fading Immunity
The signal received from any undesired path is a delayed version of the DS/SS
signal.
DS/SS signal has a property of low autocorrelation (small similarity) with its
delayed version, especially if the delay is of more than one chip duration.
The delayed signal, looking more like an interfering signal, will not be despread
by c(t) effectively minimizes the effect of the multipath signals.
What is more interesting is that DS/SS cannot only mitigate but may also
exploit the multipath propagation effect.
This is accomplished by a rake receiver.
This receive designed as to coherently combine the energy from several
multipath components, which increases the received signal power and thus
provides a form of diversity reception.
The rake receiver consists of a bank of correlation receivers, with each individual
receiver correlating with a different arriving multipath component.
By adjusting the delays, the individual multipath components can be made to add
coherently rather than destructively.
17
ECS455: Chapter 4
Multiple Access
4.5 m-sequence
Office Hours:
BKD 3601-7
Monday 9:20-10:20
Dr.Prapun Suksompong Wednesday 9:20-10:20
1 prapun.com/ecs455
Binary Random Sequences
While DSSS chip sequences must be generated
deterministically, properties of binary random sequences are
useful to gain insight into deterministic sequence design.
A random binary chip sequences consists of i.i.d. bit values
with probability one half for a one or a zero.
Also known as Bernoulli sequences/trials, “coin-flipping”
sequences
A random sequence of length N can be generated, for
example, by flipping a fair coin N times and then setting the
bit to a one for heads and a zero for tails.
2
Binary Random Sequence
X-4 X-3 X-2 X-1 X-0 X1 X2 X3 X4
Coin-flipping sequence H H T H H T H T T
Bernoulli trials/sequence 1 1 0 1 1 0 1 0 0
Binary (indp.) random sequence -1 -1 1 -1 -1 1 -1 1 1
These names are simply many versions of the same
sequence/process.
You should be able to convert one version to others
easily.
Sompe properties are conveniently explained when the
sequence is expressed in a particular version.
3
Properties of Binary Random
Sequences
4
Note: A run is a subsequence of identical symbols within the sequence.
Key randomness properties
[Golomb, 1967][Viterbi, 1995, p. 12] Binary random sequences with
length N asymptotically large have a number of the properties
desired in spreading codes
Balanced property: Equal number of ones and zeros.
Should have no DC component to avoid a spectral spike at DC or
biasing the noise in despreading
Run length property: The run length is generally short.
half of all runs are of length 1
a fraction 1/2n of all runs are of length n (Geometric)
Long runs reduce the BW spreading and its advantages
Shift property: If they are shifted by any nonzero number of
elements, the resulting sequence will have half its elements the
same as in the original sequence, and half its elements different
from the original sequence.
5 [Goldsmith,2005, p. 387 & Viterbi, p. 12]
Pseudorandom Sequence
A deterministic sequence that has the balanced, run length,
and shift properties as it grows asymptotically large is referred
to as a pseudorandom sequence (noiselike or
pseudonoise (PN) signal).
Ideally, one would prefer a random binary sequence as the
spreading sequence.
However, practical synchronization requirements in the
receiver force one to use periodic Pseudorandom binary
sequences.
m-sequences Quaternary sequences
Gold codes Walsh functions
Kasami sequences
6
m-Sequences Longer name: Maximal length
Maximal-length sequences linear shift register sequence.
A type of cyclic code
Generated and characterized by a generator polynomial
Properties can be derived using algebraic coding theory
[Goldsmith, 2005, p 387]
Simple to generate with linear feedback shift-register
(LFSR) circuits
Automated
Approximate a random binary sequence.
Disadvantage: Relatively easy to intercept and regenerate by
an unintended receiver [Ziemer, 2007, p 11]
7
(See Section 13.4.1 in [Lathi, 1998])
m-sequence generator (1)
Start with a “primitive polynomial”
The feedback taps in the feedback shift register are selected
to correspond to the coefficients of the primitive polynomial.
CLK
(Degree: r = 3 use 3 flip-flops)
1 0 x 1x 2 1x3
The gi’s are coefficients of a primitive polynomial.
1 signifies closed or a connection and
8 0 signifies open or no connection.
GF(2)
Galois field (finite field) of two elements
Consist of
the symbols 0 and 1 and
the (binary) operations of
modulo-2 addition (XOR) and
modulo-2 multiplication.
The operations are defined by
9
(See Section 13.4.1 in [Lathi, 1998])
m-sequence generator (2)
Binary sequences drawn from the alphabet {0,1} are shifted through the
shift register in response to clock pulses.
Each clock time, the register shifts all its contents to the right.
The particular 1s and 0s occupying the shift register stages after a clock
pulse are called states.
CLK
(Degree: r = 3)
1 0 x 1x 2 1x3
The gi’s are coefficients of a primitive polynomial.
1 signifies closed or a connection and
10 0 signifies open or no connection.
State Diagram
11
Primitive Polynomial
Definition: A LFSR generates an m-sequence if and only
if (starting with any nonzero state,) it visits all possible
nonzero states (in one cycle).
One can define primitive polynomial using concepts from
finite field theory.
Fact: A polynomial generates m-sequence if and only if it is a
primitive polynomial.
We then use this fact to define primitive polynomial.
For us, a polynomial is primitive if the corresponding LFSR
circuit generates m-sequence.
12
Sample Exam Question
Draw the complete state diagrams for linear feedback shift
registers (LFSRs) using the following polynomials. Does either
LFSR generate an m-sequence?
1. x3 x 2 1
2. x3 x 2 x 1
13
Nonmaximal linear feedback shift
register
x3 x 2 x 1
[Torrieri , 2005, Fig 2.8]
14
m-Sequences: More properties
1. The contents of the shift register will cycle over all possible 2r-1 nonzero states
before repeating.
2. Contain one more 1 than 0 (Slightly unbalanced)
3. Shift-and-add property: Sum of two (cyclic-)shifted m-sequences is
another (cyclic-)shift of the same m-sequence
4. If a window of width r is slid along an m-sequence for N = 2r-1 shifts, each r-
tuple except the all-zeros r-tuple will appear exactly once
5. For any m-sequence, there are
One run of ones of length r
One run of zeros of length r-1
One run of ones and one run of zeroes of length r-2
Two runs of ones and two runs of zeros of length r-3
Four runs of ones and four runs of zeros of length r-4
…
2r-3 runs of ones and 2r-3 runs of zeros of length 1
15 [S.W. Golomb, Shift Register Sequences, Holden-Day, San Francisco, 1967.]
Ex: Properties of m-sequence
00101110010111001011100101110010111001011100101110010111
Runs: 0 phase shift: 0010111
111 1 phase shift: 0101110
00 2 phase shift: 1011100
= 1100101
1,0 3 phase shift: 0111001
4 phase shift: 1110010
5 phase shift: 1100101
6 phase shift: 1001011
00101110010111001011100101110010111001011100101110010111
16
Ex: Properties of m-sequence (con’t)
25-1 = 31-chip m-sequence
1010111011000111110011010010000
1010111011000111110011010010000
Runs:
11111 1
0000 1
111 1
000 1
11 2
00 2
1 4
0 4
There are 16 runs.
17
m-Sequences (con’t)
00101110010111001011100101110010111001011100101110010111
0010111
1001011
In actual transmission, we will map 0 and 1 to +1 and -1, respectively.
Autocorrelation:
-1 1 -1 -1 -1 1 1
1 1 -1 1 -1 -1 -1
-1 1 1 -1 1 -1 -1 = -1
18
Autocorrelation and PSD
(Normalized) autocorrelations of maximal sequence and
random binary sequence.
[Torrieri , 2005, Fig 2.9]
Power spectral density of maximal sequence.
19 [Torrieri , 2005, Fig 2.10]
References: m-sequences
Karim and Sarraf, W-CDMA and
cdma2000 for 3G Mobile Networks, [TK5103.452 K37 2002]
2002.
Page 84-90
Viterbi, CDMA: Principles of Spread
Spectrum Communication, 1995
Chapter 1 and 2
Goldsmith, Wireless Communications,
2005 [TK5103.45 V57 1995]
Chapter 13
Tse and Viswanath, Fundamentals of
Wireless Communication, 2005
Section 3.4.3
20
ECS455: Chapter 4
Multiple Access
4.6 SSMA and CDMA
Office Hours:
Dr.Prapun Suksompong Library (Rangsit) Mon 16:20-16:50
www.prapun.com BKD 3601-7 Wed 9:20-11:20
1
DSSS and m-sequences
m-sequences
Excellent auto-correlation properties (for ISI rejection)
Highly suboptimal for exploiting the multiuser capabilities of
spread spectrum.
There are only a small number of maximal length codes of a
given length.
Moreover, maximal length codes generally have relatively poor
cross-correlation properties, at least for some sets of codes.
2 [Goldsmith, 2005, Ch 13]
Number of primitive polynomials
Number of different primitive
polynomials:
r is the degree of the
primitive polynomials and
Np is the number of
different primitive
polynomials available.
3 [Chen, 2007, p 145]
SSMA
For spread spectrum systems with multiple users, codes
such as Gold, Kasami, or Walsh codes are used instead of
maximal length codes
Superior cross-correlation properties.
Worse auto-correlation than maximal-length codes.
The autocorrelation function of the spreading code determines
its multipath rejection properties.
4
Qualcomm
Founders: Two of the most eminent engineers in the
world of mobile radio
Prof. Irwin Jacobs is the chairman and founder
Cornell (undergrad.: Hotel > EE)
MIT (grad.)
UCSD (Prof.)
Prof. Andrew J. Viterbi is the co-founder
MIT (BS, MS)
USC (PhD)
UCLA and UCSD (Prof.)
Same person that invented the Viterbi algorithm for
decoding convolutionally encoded data.
5
Video: Irwin Jacobs
Irwin Jacobs: Pioneer of the Wireless Future
• Gallager’s remark on ideal engineer: 1:46-2:24
• Educational background: 5:00-8:25
• Textbook: 9:03-10:40
• Viterbi: 11:00
• CDMA:
• 26:14-26:50
• 28:46-31:20
6
http://www.youtube.com/watch?v=EGaG1S4-D6o
Code Division Multiple Access (CDMA)
1991: Qualcomm announced
that it had invented a new cellular system based on
CDMA
that the capacity of this system was 20 or so times
greater than any other cellular system in existence
However, not all of the world was particularly
pleased by this apparent breakthrough—in
particular, GSM manufacturers became concerned
that they would start to lose market share to this
new system.
The result was continual and vociferous argument
between Qualcomm and the GSM manufacturers.
7
CDMA
One way to achieve SSMA
May utilize Direct Sequence Spread Spectrum (DS/SS)
The narrowband message signal is multiplied (modulated) by the spreading
signal which has a very large bandwidth (orders of magnitudes greater than
the data rate of the message).
Direct sequence is not the only spread-spectrum signaling format suitable for
CDMA Not to be confused with
error-correcting codes
All users use the same carrier frequency and may transmit that add redundancy to
combat channel noise
simultaneously. and distortion
Users are assigned different “signature waveforms” or “code” or
“codeword” or “spreading signal”
Each user’s codeword is approximately orthogonal to all other
codewords.
Should not be confused with the mobile phone standards called
cdmaOne (Qualcomm’s IS-95) and CDMA2000 (Qualcomm’s IS-2000)
(which are often referred to as simply "CDMA")
These standards use CDMA as an underlying channel access method.
8
Inner Product (Cross Correlation)
Complex conjugate
Vector *
x1 y1
n
x, y x y
*
xk yk*
k 1
x y
n n
Waveform: Time-Domain
x, y x t y* t dt
Waveform: Frequency Domain
X ,Y X f Y * f df
9
Orthogonality
Two signals are said to be orthogonal if their inner
product is zero.
The symbol is used to denote orthogonality.
* Example:
Vector: a1 b1 17
2t 3 and 5t 2 t on 1,1
n
a, b a b
*
ak bk* 0 9
k 1
a b
Time-domain: n n
a, b a t b *
t dt 0
Example (Fourier Series):
Frequency domain:
t t
sin 2 k1 and cos 2 k2 on 0, T
A, B A f B* f df 0 T T
j 2 n
t
10
e T
on 0, T
Important Properties
Parseval’s theorem
x, y x t y* t dt X f Y * f df X ,Y
If x t y t , then X f Y f .
If the non-zero regions of two signals
do not overlap in time domain or
do not overlap in frequency domain,
Then the two signals are orthogonal (their inner product = 0).
11
CDMA
Orthogonal signaling no inter-channel interference
Special cases:
TDMA: The waveforms do not overlap in the time domain.
FDMA: The waveforms do not overlap in the frequency
domain.
Orthogonal signals may overlap both in time and in
frequency domain.
12
Example: Orthogonality
An example of four “mutually orthogonal” digital signals.
c1 t When i j,
ci t , c j t 0
c2 t
c3 t
c4 t
13
Orthogonality-Based MA
CDMA
1 1
s t Sk ck t
S f S k Ck f where ck1 ck2
k 0 k 0
TDMA
1 1
s t Sk c t kTs
S f C f S k e j 2 fkTs
k 0 k 0
where c(t) is time-limited to [0,T].
This is a special case of CDMA with ck t c t kTs
The ck are non-overlapping in time domain.
FDMA 1
S f Sk C f k f
k 0
where C(f) is frequency-limited to [0,f].
This is a special case of CDMA with Ck f C f k f
14 The Ck are non-overlapping in freq. domain.
Example 1
c1 t
|C1(f)|
1
A 0.8
0.6
t 0.4
T 0.2
0
-5 -4 -3 -2 -1 0 1 2 3 4 5
f [Hz]
|C2(f)|
c2 t
1
B 0.8
0.6
0.4
t 0.2
T 0
-5 -4 -3 -2 -1 0 1 2 3 4 5
f [Hz]
-B [CDMAEx.m]
The two waveforms above overlaps both in time domain and in frequency domian.
15
Example 1 (Con’t)
Here, we use A = B. It is easy to decode the original waveforms from the shape
of the sum.
c1 t
t
T
c2 t
t 11 01 10 11 11 00 00 10
T
[Figure 1.6, Verdu, 1998]
16
Example 2: DS-CDMA
c1 t Digital version
c1 1 1 1 1
c2 t c 2 1 1 1 1
c3 1 1 1 1
c3 t c 4 1 1 1 1
1
c4 t s Sk c k
k 0
1
s t Sk ck t
17 k 0 MATLAB: ECS455_CDMAEx1.m
Block Matrix Multiplications
2 2 5 10 2 10 2 5
3 3 4 5 10 5 3 6
10 6 6 4 3
× 3 3 4 1 1 5 5 6
9 7 3 5 9
7 2 5 3 10 6 10 3
8 3 6 9 8 3 6 5
108 73 136 175 150 193 126 149
=
155 85 164 224 213 197 158 165
2 2 5 10 2 10 2 5
3 3 4 5 10 5 3 6
10 6 6 4 3
× 3 3 4 1 1 5 5 6
9 7 3 5 9
7 2 5 3 10 6 10 3
8 3 6 9 8 3 6 5
108 73 136 175 150 193 126 149
=
155 85 164 224 213 197 158 165
18
Block Matrix Multiplications
2
3 C23 5 10
4 5
2
10
D105 2
3
5
6
10 6
9 7
A B
6 4 3
3 5 9
× 3 3 4 1 1 5 5 6
7 E 2 5 3 10 F 6 10 3
8 3 6 9 8 3 6 5
108 73 136 175 150 193 126 149
=
155 85 164 224 213 197 158 165
AC+BE AD+BF
2 2 5 10 2 10 2 5
3 3 4 5 10 5 3 6
10 6 6 4 3
9 7 X3 5 9
× 3 3G 4 1 1 5 H 5 6
7 2 5 3 10 6 10 3
8 3 6 9 8 3 6 5
108 73 136 175 150 193 126 149
=
155 85 164 224 213 197 158 165
19 XG XH
Block Matrix Multiplications
c c2 cn x x1c1 x2c2 xncn where c j is m 1 and x is n 1 .
1
m n
r1
r1 x
r2 r x
x where r i is 1 m and x is m1 .
2
rn nm r n x
r1
r2
x x1 r1 x2 r 2 xn r n where r i is 1 m and x is 1 n .
rn nm
20
CDMA: DS/SS
The receiver performs a time correlation operation to detect
only the specific desired codeword.
All other codewords appear as noise due to decorrelation.
For detection of the message signal, the receiver needs to know
the codeword used by the transmitter.
Each user operates independently with no knowledge of
the other users.
Unlike TDMA or FDMA, CDMA has a soft capacity limit.
Increasing the number of users in a CDMA system raises the noise
floor in a linear manner.
There is no absolute limit on the number of users in CDMA. Rather,
the system performance gradually degrades for all users as the
number of users is increased and improves as the number of users is
decreased.
21 [Rappaport, 2002, Ch9]
Analogy [Tanenbaum, 2003]
An airport lounge with many pairs of people conversing.
TDMA is comparable to all the people being in the middle of the
room but taking turns speaking.
FDMA is comparable to the people being in widely separated
clumps, each clump holding its own conversation at the same time
as, but still independent of, the others.
CDMA is comparable to everybody being in the middle of the
room talking at once, but with each pair in a different language.
The French-speaking couple just hones in on the French, rejecting
everything that is not French as noise.
Thus, the key to CDMA is to be able to extract the desired signal
while rejecting everything else as random noise.
22
23
CDMA: Near-Far Problem
At first, CDMA did not appear to be suitable for mobile
communication systems because of this problem.
Occur when many mobile users share the same channel.
In an uplink, the signals received from each user at the receiver travel
through different channels.
Users that are close to the BS can cause a great deal of
interference to user’s farther away.
In general, the strongest received mobile signal will capture the
demodulator at a base station.
Stronger received signal levels raise the noise floor at the base station
demodulators for the weaker signals, thereby decreasing the probability
that weaker signals will be received.
Fast power control mechanisms solve this problem.
Regulate the transmit power of individual terminals in a manner that
received power levels are balanced at the BS.
24
How many orthogonal signals?
No signal can be both strictly time-limited and strictly band-
limited.
We adopt a softer definition of bandwidth and/or duration (e.g.,
the percentage of energy outside the band [-B, B] or outside the
time interval [0, T] not exceeding a given bound .
Q: How many mutually orthogonal signals with (approximate)
duration T and (approximate) bandwidth B can be constructed?
A: About 2TB
No explicit answer in terms of T, B, and is known.
Unless the product TB is small.
A K-user orthogonal CDMA system employing antipodal
modulation at the rate of R bits per second requires bandwidth
approximately equal to
1
B RK
2
25 [Verdu, 1998, Ch1, p 7]
ECS455: Chapter 4
Multiple Access
4.7 Synchronous CDMA
Office Hours:
Dr.Prapun Suksompong Library (Rangsit) Mon 16:20-16:50
www.prapun.com BKD 3601-7 Wed 9:20-11:20
1
Synchronous CDMA Model
Timing is important for orthogonality
It is not possible to obtain orthogonal codes for asynchronous
[Goldsmith, 2005, Sec. 13.4, p. 425]
users.
Bit epochs are aligned at the receiver [Verdu, 1998, p 21]
Require
Closed-loop timing control or
Providing the transmitters with access to a common clock (such
as the Global Positioning System) [Verdu, 1998, p 21]
2
Walsh Functions [Walsh, 1923]
Used in second- (2G) and
third-generation (3G) cellular
radio systems for providing
channelization
A set of Walsh functions can
be ordered according to the
number of zero crossing
(sign changes)
[Lee and Miller, 1998, Fig. 5.1]
3
Walsh Functions (2)
Orthogonality
Application:
Once we know how to generate these Walsh functions of any order N, we can use them
in N-channel orthogonal multiplexing applications.
4
Walsh Sequences
The Walsh functions, expressed
in terms of {+1,-1} values,
form a group under the
multiplication operation
(multiplicative group).
The Walsh sequences, expressed
in terms of {0, 1} values, form
a group under modulo-2
addition (additive group).
Closure property:
𝑾𝒊 𝒕 ⋅ 𝑾𝒋 𝒕 = 𝑾𝒓 𝒕
5
𝑾𝒊 ⊕ 𝑾𝒋 = 𝑾𝒓
Abstract Algebra
A group is a set of objects G on which a binary operation “ ” has
been defined. " ": G G G (closure). The operation must also
satisfy
1. Associativity: a b c a b c
2. Identity: e G such that a G a e e a a a G
3. Inverse: a G a unique element a 1 G such that
a a 1 a 1 a e .
A group is said to be commutative (or abelian) if it also satisfies
commutativity:
a, b G , a b b a .
The group operation for a commutative group is usually represented
using the symbol “+”, and the group is sometimes said to be “additive.”
6
Walsh sequences of order 64
What’s wrong with this list?! [Lee and Miller, 1998, Table 5.2]
7
Walsh Function Generation
We can construct the Walsh functions by:
1. Using Rademacher functions
2. Using Hadamard matrices
3. Exploiting the symmetry properties of Walsh functions themselves
The Hadamard matrix is a square array of plus and minus ones,
{+1, -1}, whose rows and columns are mutually orthogonal.
If the first row and first column contain only plus ones, the matrix
is said to be in normal form.
We can replace “+1” with “0” and “-1” with “1” to express the
Hadamard matrix using the logic elements {0, 1}.
The 22 Hadamard matrix of order 2 is
1 1 0 0
𝐻2 = ≡
1 −1 0 1
8
(Named after the French mathematician Jacques Hadamard)
Hadamard matrix (1)
Caution: Some textbooks write this symbol as
. It is not the regular matrix multiplication
9
If you’d like to know more,…..
Kronecker Product
An operation on two matrices of arbitrary size
Named after German mathematician Leopold Kronecker.
If A is an m-by-n matrix and B is a p-by-q matrix, then the
Kronecker product AB is the mp-by-nq matrix
a11 B a1n B
AB .
am1 B amn B
Example 1·0 1·5 2·0 2·5 0 5 0 10
1·7 2·6 2·7 6 7 12 14
1 2 0 5 1·6 .
3 4 6 7 3·0 3·5 4·0 4·5 0 15 0 20
10
3·6 3·7 4·6 4·7 18 21 24 28
Hadamard matrix (2)
11
Hadamard matrix: Examples
In MATLAB, use
hadamard(k)
12
Two ways to get H8 from H2 and H4
H2 H4
H8 H 2 H 4 H8 H 4 H 2
13
Properties
Orthogonality:
Geometric interpretation: every two different rows represent two
perpendicular vectors
Combinatorial interpretation: every two different rows have matching
entries in exactly half of their elements and mismatched entries in the
remaining elements.
Symmetric
Closure property
The elements in the first column and the first row are all 1s. The
elements in all the other rows and columns are evenly divided
between 1 and -1.
Traceless property
14
Walsh–Hadamard Sequences
All the rows (or columns) of Hadamard matrices are Walsh
sequences if the order is N = 2t.
Rows of the Hadamard matrix are not indexed according to the
number of sign changes.
Used in synchronous CDMA
It is possible to synchronize users on the downlink, where all
signals originate from the same transmitter.
It is more challenging to synchronize users in the uplink, since
they are not co-located.
Asynchronous CDMA
15
Hadamard Matrix in MATLAB
We use the hadamard function in MATLAB to generate
Hadamard matrix.
N = 8; % Length of Walsh (Hadamard) functions
hadamardMatrix = hadamard(N)
hadamardMatrix =
1 1 1 1 1 1 1 1
1 -1 1 -1 1 -1 1 -1
1 1 -1 -1 1 1 -1 -1
1 -1 -1 1 1 -1 -1 1
1 1 1 1 -1 -1 -1 -1
1 -1 1 -1 -1 1 -1 1
1 1 -1 -1 -1 -1 1 1
1 -1 -1 1 -1 1 1 -1
The Walsh functions in the matrix are not arranged in
increasing order of their sequencies or number of zero-
crossings (i.e. 'sequency order') .
16
Walsh Matrix in MATLAB
The Walsh matrix, which contains the Walsh functions along the rows or columns in the
increasing order of their sequencies is obtained by changing the index of the
hadamardMatrix as follows.
HadIdx = 0:N-1; % Hadamard index
M = log2(N)+1; % Number of bits to represent the index
Each column of the sequency index (in binary format) is given by the modulo-2 addition
of columns of the bit-reversed Hadamard index (in binary format).
binHadIdx = fliplr(dec2bin(HadIdx,M)); % Bit reversing of the binary index
binHadIdx = uint8(binHadIdx)-uint8('0'); % Convert from char to integer array
binSeqIdx = zeros(N,M-1,'uint8'); % Pre-allocate memory
for k = M:-1:2
% Binary sequency index
binSeqIdx(:,k) = xor(binHadIdx(:,k),binHadIdx(:,k-1));
end
SeqIdx = bin2dec(int2str(binSeqIdx)); % Binary to integer sequency index
walshMatrix = hadamardMatrix(SeqIdx+1,:) % 1-based indexing
walshMatrix =
1 1 1 1 1 1 1 1
1 1 1 1 -1 -1 -1 -1
1 1 -1 -1 -1 -1 1 1
1 1 -1 -1 1 1 -1 -1
1 -1 -1 1 1 -1 -1 1
1 -1 -1 1 -1 1 1 -1
1 -1 1 -1 -1 1 -1 1
17 1 -1 1 -1 1 -1 1 -1
CDMA via Hadamard Matrix
N = 8; % 8 Users
H = hadamard(N); % Hadamard matrix
%% At transmitter(s),
S = [8 0 12 0 18 0 0 10];
r = S*H
% r = 8.*H(1,:) + 12.*H(3,:) + 18.*H(5,:) + 10.*H(8,:);
% Alternatively, use
% r = ifwht(S,N,'hadamard')
%% At Receiver,
S_hat = (1/N)*r*H'
% Alternatively, use
% S_hat = fwht(r,N,'hadamard')
Specify the order of the Walsh-Hadamard
transform coefficients. ORDERING can
be 'sequency', 'hadamard' or 'dyadic'.
Default ORDERING type is 'sequency'.
Discrete Walsh-Hadamard transform
18
ECS455: Chapter 4
Multiple Access
4.8 IS-95
Office Hours:
Dr.Prapun Suksompong Library (Rangsit) Mon 16:20-16:50
www.prapun.com BKD 3601-7 Wed 9:20-11:20
1
Evolution of cellular network
[Abu-Rgheff, 2007]
2
IS-95 System
Based on direct sequence CDMA (DS-CDMA)
First CDMA-based digital cellular standard.
The brand name for IS-95 is cdmaOne.
Also known as TIA-EIA-95.
Proposed by Qualcomm in 1989 and adopted in 1993.
North America
Replaced by IS-2000 (CDMA2000)
1.25 MHz Channel BW
1.228 Mb/s chip rate
Walsh functions of “order 64” are extensively used in the
IS-95 system.
Remarks
IS-95B = cdmaOne
Upgrade IS-95A
Can carry data at rates up to 14.4 kbps for IS-95A and
115 kbps for IS-95B.
3
64-ary Walsh Functions
[Lee and Miller, 1998, Table 5.8]
4
Walsh Sequences in IS-95
Forward link (Downlink)
QPSK with a chip rate of 1,228,800 per second.
The multiple access scheme is accomplished by the use of
64-bit spreading orthogonal Walsh sequences (functions).
The (coded and interleaved) traffic channel signal symbols are multiplied
with distinct repeating Walsh sequences that are assigned to each channel
for the duration of the call.
Every base stations is synchronized with a GPS receiver so transmissions are
tightly controlled in time.
Reverse link (Uplink)
The Walsh sequences are employed as an orthogonal
modulation code, which depends only on the data pattern
(not channel), forming a 64-ary orthogonal modulation system.
5
IS-95 base station transceiver
(to provide privacy)
6
IS-95 terminal station transceiver
(for FEC)
7
IS-95
The reverse link is subject to near-far effects.
More powerful error correction is employed on the reverse link.
A rate 1/2 constraint length 9 convolutional code followed by an interleaver
on the forward channel
A rate 1/3 constraint length 9 convolutional code followed by an interleaver
is used on the reverse link.
Also with WH(6,64)
Interleaving is utilized to avoid large burst errors, which can be very
detrimental to convolutional codes.
Power control.
Use a subchannel on the forward link
Every 1.25 ms the base station receiver estimates the signal strength of the
mobile unit.
If it is too high, the base transmits a 1 on the subchannel. If it is too low, it
transmits a 0.
In this way, the mobile station adjusts its power every 1.25 ms as necessary
so as to reduce interference to other users.
8
IS-95: Increased Spectral Efficiency
Improve frequency reuse.
Narrow-band systems cannot use the same transmission frequency in
adjacent cells because of the potential for interference.
CDMA has inherent resistance to interference.
N = 1 (theoretically)
Although users from adjacent cells will contribute to interference level, their
contribution will be significantly less than the interference from the same cell
users.
Frequency reuse efficiency increases by a factor of 4 to 6.
When used to transmit voice signals, CDMA systems may exploit
the fact that voice activity typically lies at somewhat less than
40%, thus reducing the amount of interference to 40% of its
original value.
9
QCELP
Qualcomm code-excited linear prediction algorithm
Used for voice encoding.
The voice coder exploits gaps and pauses in speech.
The data rate is variable.
To keep the symbol rate constant, whenever the bit rate falls below the
peak bit rate of 9600 kbit/s, repetition is used to fill the gaps.
For example, if the output of the voice coder (and subsequently the
convolutional coder) falls to 2400 bit/s, the output is repeated three times
before it is sent to the interleaver.
Takes advantage of this repetition time by reducing the output power during
three out of the four identical symbols by at least 20 dB.
In this way, the multiple-access interference is reduced.
This voice activity gating reduces interference and increases overall
capacity.
10
References
J. S. Lee and L. E. Miller, CDMA Systems Engineering
Handbook, 1998.
Chapter 4 and 5
R.E. Ziemer, Fundamentals of Spread Spectrum
Modulation, 2007
Chapter 4
11 [TK5103.45 L44 1998]
12
ECS455: Chapter 4
Multiple Access
4.9 Async. CDMA: Gold codes and GPS
Office Hours:
Dr.Prapun Suksompong Library (Rangsit) Mon 16:20-16:50
www.prapun.com BKD 3601-7 Wed 9:20-11:20
1
Asynchronous CDMA Model
In cellular systems, the design of the reverse link (mobile-
to-base station) is considerably simplified if the users need
not be synchronized.
It is possible to let the users transmit asynchronously in
CDMA.
Codes assigned to different users need to have low cross
correlation with each other independent of the
relative delays
Gold codes
2
Gold codes
Gold codes have worse autocorrelation properties than
maximal-length codes, but better cross-correlation
properties if properly designed.
The chip sequences associated with a Gold code are produced
by addition of two m-sequences.
3
[Ziemer, 2007, Fig. 9]
Orthogonality (a revisit)
Downlinks
May use orthogonal spreading codes such as Walsh-Hadamard codes
Orthogonality can be degraded by multipath fading.
Uplinks
Generally use non-orthogonal codes due to the difficulty of user
synchronization and the complexity of maintaining code orthogonality in
uplinks with multipath.
Little dynamic coordination of users in time or frequency is required
Users can be separated by the code properties alone.
There is a hard limit on how many orthogonal channels (orthogonal
codes) can be obtained.
For non-orthogonal codes, there is no hard limit.
Non-orthogonal codes cause mutual interference between users.
The more users, the higher the level of interference
Degrade the performance of all the users.
Non-orthogonal CDMA scheme also requires power control in the
uplink to compensate for the near-far effect.
4 [Goldsmith, p 458]
Review: Near-far Effect
Arise in the uplink because the channel gain between a user’s
transmitter and the receiver is different for different users.
Suppose that one user is very close to his base station or access
point, and another user very far away.
If both users transmit at the same power level, then the interference
from the close user will swamp the signal from the far user.
Power control
Make the received signal power of all users to be roughly the same
Essentially inverts any attenuation and/or fading on the channel
Each interferer must contribute an equal amount of power
Eliminating the near-far effect
5
Global Positioning System (GPS)
Original application in the (US) military
Created in the early 1990s.
Allow a person to determine the time and the person's
precise location (latitude, longitude, and altitude) anywhere
on earth.
6
Applications
The potential applications of GPS are so vast that it has been
called (with some exaggeration) the next utility (similar to
gas, water, and electricity).
Most people probably think of it as the satellite system that
allows their satnavs to work,
but GPS is everywhere these days:
it automatically opens train doors at stations;
it tags our photos so we’ll remember where we took them;
it even keeps serves’ clocks in sync.
Its main use, however, is in preventing marital arguments on
long car journeys.
7 [PC Pro, 2011-08, 100 Technologies That Changed The World]
GPS Satellites
A minimum of 24 GPS satellites are in orbit at 20,200
kilometers (12,600 miles) above the Earth.
The satellites are spaced so that from any point on Earth, at
least four satellites will be above the horizon.
8
GPS and Gold codes
Gold codes are used to distinguish the signals from different
satellites
Coarse Acquisition Code (C/A)
Standard Positioning Service (SPS)
The message data is transmitted at 50 bits per second.
1023 bits with a period of one millisecond.
Positions of these feedback
determine the satellite ID
9
[Plausinaitis]
Auto and cross correlation of C/A code
10
How GPS Works?
A GPS receiver measuring its distance from a group of
satellites in space which are acting as precise reference points.
All the satellites have atomic clocks of unbelievable precision on
board and are synchronized.
The satellite are continuously transmitting the information about
their location and time.
GPS receiver on the ground is in synchronism with the satellites.
Off by an (unknown) amount .
For now, assume = 0.
By measuring the propagation time, the receiver can compute
distance d from that satellite.
11
GPS-Trilateration
Intersection of three sphere narrows down the location to
just two points.
[Lathi ,1998, Fig. 9.6 ]
In practice, there are four unknowns, the coordinates in the
three-dimensional space of the user along with within the
user’s receiver.
Need a distance measurement from a fourth satellite.
12
ECS455: Chapter 5
OFDM
Office Hours:
Dr.Prapun Suksompong Library (Rangsit) Mon 16:20-16:50
www.prapun.com BKD 3601-7 Wed 9:20-11:20
1
OFDM Applications
802.11 Wi-Fi: a/g/n/ac versions
DVB-T (Digital Video Broadcasting — Terrestrial)
terrestrial digital TV broadcast system used in most of the world
outside North America
DMT (the standard form of ADSL - Asymmetric Digital
Subscriber Line)
WiMAX, LTE (OFDMA)
2
Side Note: Digital TV
Japan: Starting July 24, 2011,
the analog broadcast has ceased
and only digital broadcast is
available.
US: Since June 12, 2009, full-
power television stations
nationwide have been
broadcasting exclusively in a
digital format.
Digital Video Broadcasting
– Second Generation Terrestrial
Thailand’s Roadmap:
2555 2556 2557 2558 2559
3
OFDM: Overview (1)
Let S = (S1, S2, …, SN) contains the information symbols.
𝑆
S IFFT FFT
Inverse fast Fast Fourier
Fourier transform
transform
4
OFDM: Overview (2)
Let S = (S1, S2, …, SN) be the information symbol.
The discrete baseband OFDM modulated symbol can be
expressed as
1 N 1
2 kt
s (t ) S k exp j , 0 t Ts Some references
Some references N k 0 Ts may start with
may use different different time
N 1
1 2 kt
constant in the
Sk 10,Ts t exp j interval, e.g. [-Ts/2,
front
k 0 N Ts +Ts/2]
ck t
Note that:
1
N 1
2 kt 2 kt
Re s(t ) Re Sk cos Im Sk sin
N k 0 Ts Ts
5
Single-User OFDM
6
Motivation
First, we study the wireless channel.
There are a couple of difficult problems in communication
system over wireless channel.
Also want to achieve high data rate (throughput)
7
ECS455: Chapter 5
OFDM
5.1 Wireless Channel (A Revisit)
Office Hours:
Dr.Prapun Suksompong Library (Rangsit) Mon 16:20-16:50
www.prapun.com BKD 3601-7 Wed 9:20-11:20
8
Single Carrier Digital Transmission
Baseband:
N 1
s t sk p t kTs
k 0
1, t 0, Ts
p t 10,Ts t
0, otherwise.
Passband:
x t Re s t e j 2 fct
(a) (b)
1.2
1
1
0.8
0.6
0.8
0.4
0.6 0.2
0.4 -0.2
-0.4
0.2
-0.6
-0.8
0
-1
9 -0.2
-1 0 1 2 3 4 5 6 7 8 9 -1 0 1 2 3 4 5 6 7 8 9
Time Time
Multipath Propagation
In a wireless mobile communication system, a transmitted
signal propagating through the wireless channel often
encounters multiple reflective paths until it reaches the
receiver
We refer to this phenomenon as multipath propagation
and it causes fluctuation of the amplitude and phase of the
received signal.
We call this fluctuation multipath fading.
10
Similar Problem: Ghosting
11
Wireless Comm. and Multipath Fading
The signal received consists of a number of reflected rays,
each characterized by a different amount of attenuation and
delay. v
r t x t h t n t i x t i n t
i 0
v
h t i t i
i 0
h1 t 0.5 t 0.2 t 0.2Ts 0.3 t 0.3Ts 0.1 t 0.5Ts
h2 t 0.5 t 0.2 t 0.7Ts 0.3 t 1.5Ts 0.1 t 2.3Ts
(b)
(a) (b)
1
1 1
0.8
0.8 0.8
0.6
ISI
0.6 0.6
0.4
0.4 0.4
0.2
0.2 0.2
-0.2
0
-0.2
0
(Intersymbol Interference)
-0.2
-0.4
-0.4 -0.4
-0.6
-0.6 -0.6
-0.8
-0.8 -0.8
-1
12
-1 -1
6 7 8 9 -1 0 1 2 3 4 5 6 7 8 9
-1 0 1 2 3 4 5 6 7 8 9 0 2 4 6 8 10 12
Time
Time Time
Frequency Domain
The transmitted signal (envelope) |P(f)|
1
0.5
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
f
Channel with weak multipath |H1(f)|
1.5
1
0.5
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
f
Channel with strong multipath |H2(f)|
1.5
1
0.5
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
f
13
Observation
Delay spread causes ISI
Observation: A general rule of thumb is that a delay spread of
less than 5 or 10 times the symbol width will not be a
significant factor for ISI.
Solution: The ISI can be mitigated by reducing the symbol
rate and/or including sufficient guard times between
symbols.
14
COST 207 Channel Model
Based on channel measurements with a bandwidth of 8–
10MHz in the 900MHz band used for 2G systems such as
GSM.
15
[Fazel and Kaiser, 2008, Table 1-1]
3GPP LTE Channel Modelss
16
[Fazel and Kaiser, 2008, Table 1-3]
3GPP 6-tap typical urban (TU6)
Delay profile and frequency response of 3GPP 6-tap typical
urban (TU6) Rayleigh fading channel in 5 MHz band.
[3GPP TS 45.005 – 3GPP; Technical Specification Group GSM/EDGE Radio Access
17
Network; Radio Transmission and Reception (Release 7)]
Equalization
Chapter 11 of [Goldsmith, 2005]
In a broad sense, equalization defines any signal processing
technique used at the receiver to alleviate the ISI problem
caused by delay spread. [Goldsmith, 2005]
Higher data rate applications are more sensitive to delay
spread, and generally require high-performance equalizers or
other ISI mitigation techniques.
Signal processing can also be used at the transmitter to make
the signal less susceptible to delay spread.
Ex. spread spectrum and multicarrier modulation
18
Equalizer design
Need to balance ISI mitigation with noise enhancement
Both the signal and the noise pass through the equalizer
Nonlinear equalizers suffer less from noise enhancement than
linear equalizers, but typically entail higher complexity.
Most equalizers are implemented digitally after A/D conversion
Such filters are small, cheap, easily tuneable, and very power efficient.
The optimal equalization technique is maximum likelihood
sequence estimation (MLSE).
Unfortunately, the complexity of this technique (even when using
Viterbi algorithm) grows exponentially with the length of the
delay spread, and is therefore impractical on most channels of
interest.
19
Simple Analog Equalizer
x t H eq f
H f x t n t
1
n t
Hf
Attempt to remove all ISI
Disadvantages:
If some frequencies in the channel frequency response H( f ) are
greatly attenuated, the equalizer Heq(f ) = 1/ H( f ) will greatly
enhance the noise power at those frequencies.
If the channel frequency response H( f ) has a spectral null (= 0 for
some frequency), then the power of the new noise is infinite.
Even though the ISI effects are (completely) removed, the
equalized system will perform poorly due to its greatly reduced
SNR.
20
Linear vs. Non-linear Equalizers
Linear digital equalizers
In general work by inverting the channel frequency response
Easy to implement and to understand conceptually
Typically suffer from more noise enhancement
Not used in most wireless applications
Nonlinear equalizers
Do not invert the channel frequency response
Suffer much less from noise enhancement
Decision-feedback equalization (DFE) is the most
common
Fairly simple to implement and generally performs well.
21
Equalizer Types
Symbol-by-symbol
(SBS) equalizers:
remove ISI from each Sequence estimators
symbol and then (SE): detect sequences of
detect each symbol symbols, so the effect of
individually. ISI is part of the
estimation process.
[Goldsmith, 2005, Fig. 11.2]
22
Transversal Structure
Linear and nonlinear equalizers are typically implemented
using a transversal or lattice structure.
The transversal structure is a filter with N − 1 delay
elements and N taps with tunable complex weights.
L
H eq z i
w z
i L
i
N 2L 1
The length of the equalizer N is typically dictated by
implementation considerations
Large N usually entails higher complexity.
23
Time-varying Multipath Channel
L 1
Impulse Response: h , t i t i
i 0
L = number of resolvable paths
i (t) = complex-valued path gain of the ith path
Usually assumed to be independent complex Gaussian processes resulting
in Rayleigh fading because each resolvable path is the contribution of a
different group of many irresolvable paths.
i = time delay of the ith path
Transfer function:H f , t
L = 16-path exponential power delay profile with a
decay factor of 1.0 dB and a time delay separation of
150 ns between adjacent paths (corresponding to the
rms delay spread of 0.52 μs). 5 GHz carrier
frequency and 4 km/h terminal speed.
24
[Adachi, Garg, Takaoka, and Takeda, 2005, Figure 2]
Adaptive Equalization
Equalizers must typically have an estimate of the channel
(impulse or frequency response)
Since the wireless channel varies over time, the equalizer must
learn the frequency or impulse response of the channel (training)
and then update its estimate of the frequency response as the channel
changes
The process of equalizer training and tracking is often
referred to as adaptive equalization.
Blind equalizers do not use training
Learn the channel response via the detected data only
25
Equalization for Digital Cellular
Telephony
GSM
Use adaptive equalizer
Equalize echos up to 16 ms after the first signal received
Correspond to 4.8 km in distance.
One bit period is 3.69 ms. Hence, echos with about 4 bit lengths delay
can be compensated
The direct sequence spreading employed by CDMA (IS-95)
obviates the need for a traditional equalizer.
If the transmission bandwidth is large (for example 20 MHz),
the complexity of straightforward high-performance
equalization starts to become a serious issue.
26
Wireless Propagation
[Bahai, 2002, Fig. 2.1]
27
Three steps towards modern OFDM
1. To mitigate multipath problem
Use multicarrier modulation (FDM)
2. To gain spectral efficiency
Use orthogonality of the carriers
3. To achieve efficient implementation
Use FFT and IFFT
28
ECS455: Chapter 5
OFDM
5.2 Multi-Carrier Transmission
Office Hours:
Dr.Prapun Suksompong Library (Rangsit) Mon 16:20-16:50
www.prapun.com BKD 3601-7 Wed 9:20-11:20
29
Single-Carrier Transmission
[Karim and Sarraf, 2002, Fig 3-1]
In-phase component Quadrature component
s t S I pI t cos ct SQ pQ t sin ct
Re S I pI t jSQ pQ t e jct
30
Multi-Carrier Transmission
Convert a serial high rate data stream on to multiple
parallel low rate sub-streams.
Each sub-stream is modulated on its own sub-carrier.
Time domain perspective: Since the symbol rate on each
sub-carrier is much less than the initial serial data symbol
rate, the effects of delay spread, i.e. ISI, significantly
decrease, reducing the complexity of the equalizer.
31 [Fazel and Kaiser, 2008, Fig 1-4]
Frequency Division Multiplexing
Frequency Domain Perspective: Even though the fast
fading is frequency-selective across the entire OFDM signal
band, it is effectively flat in the band of each low-speed
signal. [The flatness assumption is the same one that you used
in Riemann approximation of integral.]
32 [Myung and Goodman, 2008]
Frequency Division Multiplexing
To facilitate separation of the signals at the receiver, the
carrier frequencies were spaced sufficiently far apart so
that the signal spectra did not overlap. Empty spectral
regions between the signals assured that they could be
separated with readily realizable filters.
The resulting spectral efficiency was therefore quite low.
33
Multi-Carrier (FDM) vs. Single Carrier
Single Carrier Multi-Carrier (FDM)
Single higher rate serial scheme Parallel scheme. Each of the parallel
subchannels can carry a low signalling rate,
proportional to its bandwidth.
Multipath problem: Far more susceptible Long duration signal elements and
to inter-symbol interference (ISI) due to the narrow bandwidth in sub-channels.
short duration of its signal elements and the Complexity problem: If built
higher distortion produced by its wider straightforwardly as several (N) transmitters
frequency band and receivers, will be more costly to
Complicated equalization implement.
BW efficiency problem: The sum of
parallel signalling rates is less than can be
carried by a single serial channel of that
combined bandwidth because of the unused
guard space between the parallel sub-
carriers.
34
FDM (con’t)
Before the development of equalization, the parallel
technique was the preferred means of achieving high rates
over a dispersive channel, in spite of its high cost and relative
bandwidth inefficiency.
35
OFDM
OFDM = Orthogonal frequency division multiplexing
One of multi-carrier modulation (MCM) techniques
Parallel data transmission (of many sequential streams)
A broadband is divided into many narrow sub-channels
Frequency division multiplexing (FDM)
High spectral efficiency
The sub-channels are made orthogonal to each other over the OFDM
symbol duration Ts.
Spacing is carefully selected.
Allow the sub-channels to overlap in the frequency domain.
Allow sub-carriers to be spaced as close as theoretically possible.
36
OFDM
Recall: Orthogonality-Based MA (CDMA)
1
s t Sk ck t where ck1 ck2
k 0
Discrete baseband OFDM modulated symbol:
1 N 1
2 kt
s (t ) S k exp j , 0 t Ts
N k 0 Ts
N 1
1 2 kt
Sk 10,Ts t exp j
k 0 N Ts
ck t
Another special case of CDMA!
37
OFDM: Orthogonality
2 k1t 2 k2t
Ts
ck1 t c t dt 0 exp exp j
*
k2 j dt
Ts Ts
Ts
2 k1 k2 t Ts , k1 k2
exp j dt
0 Ts 0, k1 k2
When k1 k2 , Ts
k1 k2 t dt 1dt Ts
*
c t c
0
When k1 k2 ,
2 k1 k2 t
Ts
Ts
ck1 t c t dt j 2 k1 k2 exp j
*
k2
Ts 0
Ts
1 1 0
j 2 k1 k2
38
Frequency Spectrum
N 1
1 2 kt f
1
s (t ) Sk 10,Ts t exp j Ts
k 0 N Ts
ck t
This is the term
1 Ts Ts t Ts sin c Ts f that makes the
2 , 2 technique FDM.
Ts
1 1 j 2 f
c t 10,Ts t
C f Ts e 2
sin c Ts f
N N
2 kt k
ck t c t exp j Ck f C f C f k f
Ts Ts
N 1 N 1
s(t ) Sk ck t
S ( f ) S k Ck f
k 0 k 0
N 1 j 2 f k f
Ts
Ts sin c Ts f k f
1
39
N
Se
k 0
k
2
N 1
1 2 kt
s (t ) S k 10,Ts t exp j
k 0 N Ts
Subcarrier Spacing
N 1 j 2 f k f
Ts
Ts sin c Ts f k f
1
S f
N
S e
k 0
k
2
1
f
Ts
Each QAM signal carries one of N separate QAM signals,
OFDM FDM
the original input complex at N frequencies
numbers. separated by the signaling rate.
The spectrum of each
QAM signal is of the
form with nulls at the
center of the other sub-
Spectrum Overlap in OFDM
40 carriers.
Normalized Power Density Spectrum
More flat with more sub-carriers
http://www.metageek.net/forums/showthread.php?4912-Chanalyzer-4-User-Guide
[Fazel and Kaiser, 2008, Fig 1-5]
41
OFDM Carriers: N = 4
2 kt 2 kt
cos sin
s
T s
T
1 1
0 0
-1 -1
0 0.5 1 0 0.5 1
1 1
0 0
-1 -1
0 0.5 1 0 0.5 1
1 1
0 0
-1 -1
0 0.5 1 0 0.5 1
1 1
0 0
-1 -1
0 0.5 1 0 0.5 1
42
OFDM as a Multicarrier Technique
1
N 1
2 kt 2 kt
Re s(t ) Re Sk cos Im Sk sin
N k 0 s
T s
T
1
0
-1
0
1
0
-1
0
1
0
-1
0
1
0
-1
0
1
0
-1
0
1
0
-1
0
1
0
-1
1
0
43 -1
Time-Domain Signal
Real and Imaginary
components
of an OFDM symbol is the
superposition of several
harmonics
modulated by data symbols
[Bahai, 2002, Fig 1.7]
1 N 1
2 kt
s(t ) Sk exp j , 0 t Ts
N k 0 Ts
1 N 1 2 kt 2 kt
Re s(t ) Re Sk cos Im Sk sin
44 N k 0 s
T s
T
in-phase part quadrature part
Summary
So, we have a scheme which achieves
Large symbol duration (Ts) and hence less multipath problem
Good spectral efficiency
One more problem:
There are so many carriers!
45
Xˆ 0
4-pt DFT
x[0] = f[0] 2-pt DFT Xˆ 21 0 Xˆ 11 0 Fˆ 0
x[4] = f[2] Xˆ 21 1 Xˆ 11 1 Fˆ 1 Xˆ 1
-1
1
4
1
8
x[2] = f[1] 2-pt DFT
Xˆ 22 0 Xˆ 11 2 Fˆ 2 Xˆ 2
42 82
x[6] = f[3] Xˆ 22 1 Xˆ 11 3 Fˆ 3 Xˆ 3
-1 43 3
8
-1 DFT
Xˆ 23 0 4-pt
Xˆ 12 0 Gˆ 0 Xˆ 4
ECS455: Chapter 5
x[1] = g[0] 2-pt DFT
84
x[5] = g[2] Xˆ 23 1 Xˆ 12 1 Gˆ 1 Xˆ 5
-1 85
1
4
x[3] = g[1] 2-pt DFT Xˆ 24 0 Xˆ 12 2 Gˆ 2 Xˆ 6
42 86
x[7] = g[3]
-1 Xˆ 24 1
43
-1
OFDM Xˆ 12 3 Gˆ 3 87
Xˆ 7
5.3 Implementation: DFT and FFT
Office Hours:
Dr.Prapun Suksompong Library (Rangsit) Mon 16:20-16:50
www.prapun.com BKD 3601-7 Wed 9:20-11:20
1
Discrete Fourier Transform (DFT)
Transmitter produces
1 N 1
2 kt
s (t ) S k exp j , 0 t Ts
N k 0 Ts
Sample the signal in time domain every Ts/N gives
T 1 N 1 2 k Ts
s n s n s S k exp j n
N N
N k 0 Ts
1 N 1 2 kn
S k exp j N IDFT S n
N k 0 N
We can implement OFDM in the discrete domain!
2
Discrete Fourier Transform (DFT)
The “-1” are there because
we start from row 1 and
column 1.
Key Property:
1
N is a unitary matrix
3
N
DFT
4
DFT: Example
5 [http://www.fourier-series.com/fourierseries2/DFT_tutorial.html]
DTFT to DFT
Start with a sequence in discrete time x[n].
Z-transform: X z x n z n
n
Discrete-Time Fourier Transform: X e j x n e jn
n
N 1
N points in time domain: X e j
x n e j n
n 0
N 1 2
DFT: X k X e j
j
x n e
kn
N
k
k 2
N n 0
6
Efficient Implementation: (I)FFT
[Bahai, 2002, Fig. 2.9]
An N-point FFT requires only on the order of NlogN multiplications, rather than N2 as in
7 a straightforward computation.
FFT
The history of the FFT is complicated.
As with many discoveries and inventions, it arrived before the
(computer) world was ready for it.
Usually done with N a power of two.
Very efficient in terms of computing time
Ideally suited to the binary arithmetic of digital computers.
Ex: From the implementation point of view it is better to have, for
example, a FFT size of 1024 even if only 600 outputs are used than
try to have another length for FFT between 600 and 1024.
References: E. Oran Brigham, The Fast
Fourier Transform, Prentice-Hall, 1974.
8
1 N 1
2 kt
s (t ) S k exp j , 0 t Ts
DFT Samples N k 0 Ts
Here are the points s[n] on the continuous-time version s(t):
0.6
Ts
s n s n
N
0.4
0.2
1 N 1 2 kn
0
N k 0
S k exp j
N
-0.2 N IDFT S n
-0.4
0n N
-0.6
-0.8
0 1 2 3 4 5 6 7 8
Ts
9
Oversampling
0.6
0.4
0.2
-0.2
-0.4
-0.6
-0.8
0 1 2 3 4 5 6 7 8
10
Oversampling (2)
Increase the number of sample points from N to LN on the
interval [0,Ts].
L is called the over-sampling factor.
Ts Ts
s n s n s
L
n
n s
N LN
0n N 0 n LN
1 N 1 2 k Ts 1 N 1
2 kn
s
L
n S k exp j
n
S k exp j
N k 0 Ts
LN N k 0 LN
1 1 N 1
2 kn
N
LN
LN
S
k 0
k exp j
LN
Zero padding:
1 N 1 2 kn NL 1 2 kn
Sk , 0 k N
L N k S exp j
0 exp j Sk
LN k 0 LN kN LN
0, N k LN
1 NL 1
2 kn
11 L N S k exp j
L N IDFT S n
LN k 0 LN
Oversampling: Summary
N points LN points
Ts Ts
s n s n N IDFT S n s n s n
L
L N IDFT S n
0 n N 0 n LN LN
N
Zero padding:
Sk , 0 k N
Sk
0, N k LN
0.6 0.6
0.4 0.4
0.2 0.2
0 0
-0.2 -0.2
-0.4 -0.4
12 -0.6 -0.6
-0.8 -0.8
0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8
OFDM implementation by IFFT/FFT
n s n
Ts
s
L
LN
L
L N IFFT S n
Sk
Rk S k
This form of OFDM is often
referred to as Discrete
13 r n = N IFFT S n Multi-Tone (DMT).
OFDM with Memoryless Channel
h t t should be h t t
r t h t s t w t s t w t
Additive white Gaussian noise
Sample every Ts/N
r n s n w n
FFT s n = N IFFT S n
1 1
Rk FFT y n S k Wk
N N
Sub-channel are independent.
(No ICI)
14
Channel with Finite Memory
Discrete time baseband model:
v
y n h * s n w n h m s n m w n
m0
[Tse Viswanath, 2005, Sec. 2.2.3]
where h n 0 for n 0 and n
i .i .d . We will assume that N
w n ~ CN 0, N 0
Remarks:
15
OFDM Architecture
[Bahai, 2002, Fig 1.11]
16
ECS455: Chapter 5
OFDM
5.4 Cyclic Prefix (CP)
Office Hours:
BKD 3601-7
Wednesday 15:30-16:30
Dr.Prapun Suksompong Friday 9:30-10:30
1 prapun.com/ecs455
Three steps towards modern OFDM
1. Mitigate Multipath (ISI): Decrease the rate of the original
data stream via multicarrier modulation (FDM)
2. Gain Spectral Efficiency: Utilize orthogonality
3. Achieve Efficient Implementation: FFT and IFFT
Extra step: Completely eliminate ISI and ICI
Cyclic prefix
2
Cyclic Prefix: Motivation (1)
Recall: Multipath Fading and Delay Spread
3
Cyclic Prefix: Motivation (2)
OFDM uses large symbol duration Ts
compared to the duration of the impulse response τmax of the channel
to reduce the amount of ISI
Q: Can we “eliminate” the multipath (ISI) problem?
A: To reduce the ISI, add guard interval larger than that of the
estimated delay spread.
If the guard interval is left empty, the orthogonality of the sub-carriers
no longer holds, i.e., ICI (inter-channel interference) still exists.
Solution: To prevent both the ISI as well as the ICI, OFDM symbol is
cyclically extended into the guard interval.
4
Cyclic Prefix
5
Recall: Convolution
𝑥𝑚
Flip
m
Shift
N-1
N
Multiply (pointwise) ℎ𝑚
m
Add
v-1
v
ℎ −𝑚 = ℎ 0 − 𝑚
m
ℎ 1−𝑚
ℎ 𝑛−𝑚
m
n
x h n x m h n m
ℎ 𝑁−1−𝑚
m
m
ℎ 𝑁−𝑚
N
ℎ 𝑁+𝐿−1−𝑚
6 m
Circular Convolution (Regular Convolution)
𝑥𝑚 𝑥𝑚
m m
N-1
N-1
N
N
ℎ𝑚 ℎ𝑚
m m
v-1
v-1
v
v
ℎ −𝑚 = ℎ 0 − 𝑚 ℎ −𝑚 = ℎ 0 − 𝑚
m m
ℎ 1−𝑚 ℎ 1−𝑚
m m
ℎ 𝑛−𝑚 ℎ 𝑛−𝑚
m m
n
n
ℎ 𝑁−1−𝑚 ℎ 𝑁−1−𝑚
m m
ℎ 𝑁−𝑚
Replicate x (now it looks periodic) m
Then, perform the usual convolution
N
only on n = 0 to N-1 ℎ 𝑁+𝐿−1−𝑚
7 m
Circular Convolution: Examples 1
Find
1 2 3 4 5 6
1 2 3 4 5 6
1 2 3 0 0 4 5 6 0 0
8
Discussion
Regular convolution of an N1–point vector and an N2–point
vector gives (N1+N2-1)-point vector.
Circular convolution is performed between two equal-
length vectors. The results also has the same length.
Circular convolution can be used to find the regular
convolution by zero-padding.
Zero-pad the vectors so that their length is N1+N2-1.
Example:
1 2 3 0 0 4 5 6 0 0 1 2 3 4 5 6
In modern OFDM, we want to perform circular convolution
via regular convolution.
9
Circular Convolution in Communication
We want the receiver to obtain the circular convolution of the
signal (channel input) and the channel.
Q: Why?
A:
CTFT: convolution in time domain corresponds to
multiplication in frequency domain.
This fact does not hold for DFT.
DFT: circular convolution in (discrete) time domain corresponds
to multiplication in (discrete) frequency domain.
We want to have multiplication in frequency domain.
So, we want circular convolution and not the regular convolution.
Problem: Real channel does regular convolution.
Solution: With cyclic prefix, regular convolution can be used to
create circular convolution.
10
Let’s look closer at how
we carry out the circular
Example 2 convolution operation.
1 2 3 1 2 3 2 1 0 0 ? Recall that we replicate
the x and then perform the
Solution: regular convolution (for N
points)
1 2 3 1 2 1 2 3 1 2 1 2 3 1 2
0 0 1 2 3 1 1 2 2 1 3 1 4 3 8
0 0 1 2 3 2 1 1 2 2 3 2 2 6 2
0 0 1 2 3 1 1 2 2 3 3 1 4 9 6
0 0 1 2 3 2 1 3 2 1 3 2 6 3 7
0 0 1 2 3 3 1 1 2 2 3 3 2 6 11
1 2 3 1 2 3 2 1 0 0 8 2 6 7 11
11 Goal: Get these numbers using regular convolution
Observation: We don’t need
to replicate the x indefinitely.
Example 2 Furthermore, when h is
shorter than x, we need only
1 2 3 1 2 3 2 1 0 0 ?
a part of one replica.
Not needed in the calculation
1 2 3 1 2 1 2 3 1 2 1 2 3 1 2
0 0 1 2 3 1 1 2 2 1 3 1 4 3 8
0 0 1 2 3 2 1 1 2 2 3 2 2 6 2
0 0 1 2 3 1 1 2 2 3 3 1 4 9 6
0 0 1 2 3 2 1 3 2 1 3 2 6 3 7
0 0 1 2 3 3 1 1 2 2 3 3 2 6 11
1 2 3 1 2 3 2 1 0 0 8 2 6 7 11
12
Try this: use only the necessary part
of the replica and then convolute
Example 2 with the channel.
(regular convolution)
1 2 1 2 3 1 2 * 3 2 1 ?
Copy the last samples of the symbols at the beginning of the symbol.
This partial replica is called the cyclic prefix.
1 2 1 2 3 1 2
1 2 3 1 3 3
1 2 3 1 2 2 3 2 6 8
1 2 3 1 1 2 2 1 3 1 4 3 8
1 2 3 2 1 1 2 2 3 2 2 6 2
1 2 3 1 1 2 2 3 3 1 4 9 6
1 2 3 2 1 3 2 1 3 2 6 3 7
1 2 3 3 1 1 2 2 3 3 2 6 11
1 2 3 1 1 2 2 1 4 5
1 2 3 2 1 2
13 Junk!
Example 2
We now know that
1 2 1 2 3 1 2 * 3 2 1 3 8 8 2 6 7 11 5 2
Cyclic Prefix 1 2 3 1 2 3 2 1 0 0
Similarly, you may check that
2 1 2 1 3 2 1 * 3 2 1 6 1 6 8 5 11 4 0 1
Cyclic Prefix 2 1 3 2 1 3 2 1 0 0
14
Example 3
We know, from Example 2, that
[ 1 2 1 -2 3 1 2] * [3 2 1] = [ 3 8 8 -2 6 7 11 5 2]
And that
[-2 1 2 1 -3 -2 1] * [3 2 1] = [-6 -1 6 8 -5 -11 -4 0 1]
Check that
[ 1 2 1 -2 3 1 2 0 0 0 0 0 0 0] * [3 2 1]
= [ 3 8 8 -2 6 7 11 5 2 0 0 0 0 0 0 0]
and
[ 0 0 0 0 0 0 0 -2 1 2 1 -3 -2 1] * [3 2 1]
= [ 0 0 0 0 0 0 0 -6 -1 6 8 -5 -11 -4 0 1]
15
Example 4
We know that
[ 1 2 1 -2 3 1 2] * [3 2 1] = [ 3 8 8 -2 6 7 11 5 2]
[-2 1 2 1 -3 -2 1] * [3 2 1] = [-6 -1 6 8 -5 -11 -4 0 1]
Using Example 3, we have
[ 1 2 1 -2 3 1 2 -2 1 2 1 -3 -2 1] * [3 2 1]
= [ 1 2 1 -2 3 1 2 0 0 0 0 0 0 0]
* [3 2 1]
+[ 0 0 0 0 0 0 0 -2 1 2 1 -3 -2 1]
= [ 3 8 8 -2 6 7 11 5 2 0 0 0 0 0 0 0]
+[ 0 0 0 0 0 0 0 -6 -1 6 8 -5 -11 -4 0 1]
= [ 3 8 8 -2 6 7 11 -1 1 6 8 -5 -11 -4 0 1]
16
Putting results together…
Suppose x(1) = [1 -2 3 1 2] and x(2) = [2 1 -3 -2 1]
Suppose h = [3 2 1]
At the receiver, we want to get
[1 -2 3 1 2] * [3 2 1 0 0] = [8 -2 6 7 11]
[2 1 -3 -2 1] * [3 2 1 0 0] = [6 8 -5 -11 -4]
We transmit [ 1 2 1 -2 3 1 2 -2 1 2 1 -3 -2 1].
Cyclic prefix Cyclic prefix
At the receiver, we get
[ 1 2 1 -2 3 1 2 -2 1 2 1 -3 -2 1] * [3 2 1]
= [ 3 8 8 -2 6 7 11 -1 1 6 8 -5 -11 -4 0 1]
Junk! To be thrown away by the receiver.
17
Circular Convolution: Key Properties
Consider an N-point signal x[n]
Cyclic Prefix (CP) insertion: If x[n] is extended by
copying the last samples of the symbols at the beginning of
the symbol:
x n, 0 n N 1
x n
x n N , v n 1
Key Property 1:
h x n h * x n for 0 n N 1
Key Property 2:
h x n
FFT
Hk X k
18
OFDM with CP for Channel w/ Memory
We want to send N samples S0, S1, …, SN-1 across noisy
channel with memory.
First apply IFFT: Sk IFFT
s n
Then, add cyclic prefix
s s N , , s N 1, s 0, , s N 1
This is inputted to the channel.
The output is
y n p N , , p N 1, r 0, , r N 1
Remove cyclic prefix to get r n h n s n w n
Then apply FFT: r n Rk
FFT
19 By circular convolution property of DFT, Rk H k Sk Wk No ICI!
OFDM System Design: CP
A good ratio between the CP interval and symbol duration
should be found, so that all multipaths are resolved and not
significant amount of energy is lost due to CP.
As a thumb rule, the CP interval must be two to four times
larger than the root mean square (RMS) delay spread.
20 [Tarokh, 2009, Fig 2.9]
Summary
The CP at the beginning of each block has two main
functions.
As guard interval, it prevents contamination of a block by ISI
from the previous block.
It makes the received block appear to be periodic of period
N.
Turn regular convolution into circular convolution
Point-wise multiplication in the frequency domain
21
Reference
A. Bahai, B. R. Saltzberg, and M.
Ergen, Multi-Carrier Digital
Communications:Theory and
Applications of OFDM, 2nd ed.,
New York: Springer Verlag, 2004.
22
ECS455: Chapter 5
OFDM
5.5 Remarks about OFDM
Office Hours:
Dr.Prapun Suksompong Library (Rangsit) Mon 16:20-16:50
www.prapun.com BKD 3601-7 Wed 9:20-11:20
1
Summary: OFDM Advantages
For a given channel delay spread, the implementation
complexity is much lower than that of a conventional single
carrier (SC) system with time domain equalizer.
Spectral efficiency is high since it uses overlapping orthogonal
subcarriers in the frequency domain.
Modulation and demodulation are implemented using inverse
discrete Fourier transform (IDFT) and discrete Fourier transform
(DFT), respectively, and fast Fourier transform (FFT) algorithms
can be applied to make the overall system efficient
(computationally).
Capacity can be significantly increased by adapting the data
rate per subcarrier according to the signal-to-noise ratio
(SNR) of the individual subcarrier.
2
Example: 802.11a
1
4 0.8 s
3
OFDM Drawbacks
High peak-to-average power ratio (PAPR)
The transmitted signal is a superposition of all the subcarriers with
different carrier frequencies and high amplitude peaks occur because
of the superposition.
High sensitivity to frequency offset:
When there are frequency offsets in the subcarriers, the
orthogonality among the subcarriers breaks and it causes intercarrier
interference (ICI).
A need for an adaptive or coded scheme to overcome spectral
nulls in the channel
In the presence of a null in the channel, there is no way to recover the
data of the subcarriers that are affected by the null unless we use rate
adaptation or a coding scheme.
4
ECS455: Chapter 5
OFDM
5.6 OFDM-Based Multiple Access Techniques
Office Hours:
Dr.Prapun Suksompong Library (Rangsit) Mon 16:20-16:50
www.prapun.com BKD 3601-7 Wed 9:20-11:20
5
OFDM-based Multiple Access
1. OFDMA
2. OFDM-TDMA
3. Multi-Carrier CDMA (OFDM-CDMA)
Lowest
flexibility
Easiest to
implement
[Tarokh, 2009, Section 2.9, Fig 2.10]
6
subcarriers
OFDM-TDMA
Users are separated via time slots.
A particular user uses all sub-carriers within
the predetermined TDMA time slot.
time
Example: 802.11
Each user uses OFDM modulation and gets
transmission right through the MAC layer
channel access mechanism.
7
OFDM-TDMA (2)
Advantage:
MS can reduce its power consumption
Process only OFDM symbols which are dedicated to it.
Disadvantage:
Allocate the whole bandwidth to a single user
A reaction to different subcarrier attenuations could consist of leaving out
highly distorted subcarriers
8
OFDMA
Available subcarriers are distributed among all the users for
transmission at any time instant.
The fact that each user experiences a different radio channel
can be exploited by allocating only “good” subcarriers with
high SNR to each user.
Recall: For OFDM system, based on the subchannel
condition, different baseband modulation schemes can
be used for the individual subchannels
The number of subchannels for a specific user can be varied,
according to the required data rate.
9
OFDM-TDMA vs. OFDMA
subcarriers
subcarriers
OFDM-TDMA OFDMA
time time
10
OFDMA Block Diagram
Channel state information for all K
Combined subcarrier, bit,
users
and power allocation x N subcarriers/user
Adaptive mod. 1
User 1 data
User 2 data Subcarrier Adaptive mod. 2 Add guard
and bit IFFT interval
allocation
User K data Adaptive mod. N
Freq. selective
fading channel
for User k
Adaptive demod. 1
Adaptive demod. 2 Remove
User k data Extract bits
FFT guard
for User k interval
Adaptive demod. N
11
ECS455: Chapter 6
Applications
Office Hours:
Dr.Prapun Suksompong Library (Rangsit) Mon 16:20-16:50
www.prapun.com BKD 3601-7 Wed 9:20-11:20
1
Review: Cellular History
The first-generation (1G) systems introduced in the
1980s were characterized by analog speech
transmission.
The second-generation (2G) systems, developed in the
1990s, are digital and have data transport capabilities but
only to a limited extent.
GSM supports SMSs and user data at rates only up to 9.6 kb/s.
IS-95B (cdmaOne) provides data rates in the range of 64 to
115 kb/s in increments of 8 kb/s over a 1.25 MHz channel.
Each cell uses a carrier with a bandwidth of 1.25MHz, which is divided
into 64 data and signalling channels by the use of orthogonal CDMA
codes.
2
Review: GSM (2G)
[Karim and Sarraf, 2002, Fig 5-1]
[Karim and Sarraf, 2002, Fig 5-10]
3
GSM Enhancement
Want to deliver data as well as voice.
2.5G: General Packet Radio Service (GPRS)
Provide connectivity to IP networks (Internet).
Each slot can handle up to 20 kb/s. Each user may be allocated
up to 8 slots
Data rates up to about 160 kb/s per user are possible.
A single time slot may be shared by multiple users for
transferring packet mode data.
2.75G: Enhanced Data Rates for GSM Evolution
(EDGE)
Support IP-based services in GSM at rates up to 384 kb/s
4
ECS455: Chapter 6
Applications
6.1 3G (UMTS and WCDMA)
Office Hours:
Dr.Prapun Suksompong Library (Rangsit) Mon 16:20-16:50
www.prapun.com BKD 3601-7 Wed 9:20-11:20
5
Review: UMTS
Universal Mobile Telecommunications System (UMTS)
The research activity on UMTS started in Europe at the beginning of
the 1990s.
Even before the earliest 2G systems arrived on the market
Designed to support wideband services with data rates up to 2Mbit/s.
Developed from GSM
Keep the core network more-or-less intact
Change the air interface to use CDMA
Compatibility between UMTS and GSM:
Most UMTS mobiles also implement GSM, and the network can hand them
over from a UMTS base station to a GSM one if they reach the edge of the
UMTS coverage area.
However, network operators cannot implement the two systems in the same
frequency band, so they are not fully compatible with each other.
6
Bandwidth Comparison
Generation Transmission Bandwidth Standard
1G 25 and 30 kHz
2G 200 kHz GSM
1.25 MHz IS-95 (CDMA)
2.5G 1.25 MHz CDMA2000 1X-RTT
3G 5 MHz WCDMA and CDMA2000 3X-RTT
“4G” Up to 20 MHz LTE and WiMAX (802.16)
Wi-Fi (802.11a/g)
Wider and wider radio frequency bands!
7 [Myung and Goodman, 2008]
UMTS: FDD
The chip rate for spectrum spreading is 3.84 Mc/s.
The maximum transmitter power of the user equipment is in
the range of 21 to 33 dBm (that is, 125 mW to 2 W)
1021/10 mW 1033/10
mW
(0 dBm or dBmW = 1 mW)
[Karim and Sarraf, 2002, Fig 6-1]
8
UMTS W-CDMA FDD
(UMTS Terrestrial Radio Access)
and TDD
The air interface (UTRA) of UMTS is based on CDMA
UMTS W-CDMA FDD
Direct-sequence CDMA system
5 MHz bandwidth
UMTS W-CDMA TDD
Also uses CDMA with a bandwidth of 5 MHz
The frequency band is time shared in both directions—one half
of the time, it is used for transmission in the forward direction
and the other half of the time in the reverse direction.
FDD is currently much more popular
9
Cdma2000 (IMT Multi-Carrier)
Another 3G mobile technology standard
Multicarrier, direct-sequence CDMA FDD system.
Backward-compatible with its previous 2G iteration IS-95
(cdmaOne).
CDMA2000 1X (IS-2000)
also known as 1x and 1xRTT
1x = Spreading Rate 1 = use the same chip rate of IS-95 (i.e., 1.2288
Mcps).
Same RF bandwidth as IS-95: a duplex pair of 1.25 MHz radio channels.
Core CDMA2000 wireless air interface standard.
Almost doubles the capacity of IS-95 by adding 64 more traffic
channels to the forward link, orthogonal to (in quadrature with) the
original set of 64.
10
3G Spreading Codes
In UMTS and cdma2000, signaling and user data is spread
twice in succession
First with the channelization codes
Orthogonal Walsh codes
Inherently more tolerant of interference caused by multiple users.
Later with the scrambling codes
Not necessarily orthogonal
Built from PN codes
In contrast to IS-95, the WCDMA/UMTS standard applies
variable length orthogonal spreading codes and coherent
QPSK detection for both uplink and downlink directions.
11
Review: CDMA
Two Users. Suppose the code length = 4.
User 1 uses code c1. Want to send messages a1, a2, a3, a4, …
Send x1 a1 c1 a2 c1 a3 c1 a4 c1
User 2 uses code c2. Want to send messages b1, b2, b3, b4, …
Send x 2 b1 c 2 b2 c 2 b3 c 2 b4 c 2
Receiver gets r x1 x 2
1
To recover a1, calculate r 1: 4 , c1
4
1
To recover b1, calculate r 1: 4 , c 2
4
To recover a2, calculate 1
r 5 : 8 , c1
4
12
OVSF (1)
Channelization codes used in UMTS W-CDMA and cdma2000 are
variable-length Walsh codes, also known as orthogonal
variable spreading factor (OVSF) codes.
The spreading factors in UMTS may vary from 4 to 256 chips on
uplink channels and from 4 to 512 chips on downlink channels.
In cdma2000, OVSF codes used on traffic channels may vary from 4
to 128 chips.
Comparison: IS-95 uses a set of 64 fixed-length Walsh codes to
spread forward physical channels. In the reverse direction, they
are used for orthogonal modulation where every six symbols from
the block interleaver output are modulated as one of 64 Walsh
codes.
13
OVSF (2)
Similar to Walsh sequences
Arranged and numbered in a different way
Use a tree structure
For each spreading factor SF = 1, 2, 4, . . . , which is a
power of 2, there are N = SF orthogonal codes obtained by
the recursion relations:
Different data rates are supported on a physical channel by
simply changing the spreading factor of the associated code.
14
Tree structure for OVSF codes (1)
15
Tree structure for OVSF codes (2)
16
Code allocation rules (1)
OVSF codes can be applied to realize connections with
different data rates by varying the spreading factor.
Smaller SF = Faster data rate
To have connections with different data rates, need some
rules (for selecting the codes) to maintain orthogonality
Code blocking property: If a certain code is already used
for one connection, neither this code nor a code that is a
descendant or an ancestor of this code (on the tree) is
allowed to be used for another connection
These codes are not orthogonal to the already allocated one.
17
Code allocation rules (2)
[Schulze Luders, 2005, Fig 5.12]
18
Code allocation rules (3)
Two OVSF codes are orthogonal if and only if neither code
lies on the path from the other code to the root.
If, for example, code c4,1 is in use, another connection with a
different data rate is not allowed to use the encircled codes.
Other cades can still be used.
If, for example, the second connection has twice the data rate
of the first one, it has to select the code c2,1.
Within the period of one data bit of connection 1, connection 2
transmits two data bits.
19
OVSF: Disadvantages
Poor autocorrelation property
Look, for example, at the codes cSF,0.
When there is no perfect synchronization, the orthogonality
gets lost (high values for the cross correlation)
20
Scrambling Codes in UMTS
The scrambling codes in UMTS are complex valued and may be either long or
short.
A long code has a length of 38,400 chips (that is, 10 ms) and a short code only
256 chips.
A long code for a UMTS uplink channel is constructed with two PN codes,
whose characteristic polynomials are
g1 x x 25 x3 1 and g 2 x x 25 x3 x 2 x 1
They are implemented as sequences PN1 and PN2 using two 25-bit shift
registers.
PN1 and PN2 are added modulo 2, and the output is mapped to a real-valued
function, say, I.
Another function Q is derived by simply delaying I by 224 + 16 chips.
Q is multiplied by j, where the sign changes every chip period, and then added
to I to yield the long code.
21
You might also like
- Multiple Access Tech and DS CDMA 03 Feb 23No ratings yetMultiple Access Tech and DS CDMA 03 Feb 23154 pages
- Chapter-2 (WMC) (1) Wireless CommunicationNo ratings yetChapter-2 (WMC) (1) Wireless Communication59 pages
- Multiple Access Techniques: by - Sanket Jain - Rishabh LodhaNo ratings yetMultiple Access Techniques: by - Sanket Jain - Rishabh Lodha23 pages
- Unit-2: Cellular Concept and Multiple Access Techniques100% (1)Unit-2: Cellular Concept and Multiple Access Techniques46 pages
- Multiple Access Techniques For Wireless Communication: Duplexing Fdma Tdma Sdma PdmaNo ratings yetMultiple Access Techniques For Wireless Communication: Duplexing Fdma Tdma Sdma Pdma41 pages
- Part7 - Duplexing and Multiple Access TechniquesNo ratings yetPart7 - Duplexing and Multiple Access Techniques25 pages
- Multiple Access-FDMA, CDMA, TDMA, SDMA, DSSS, FHSS, ALOHA, PACKET RADIO...100% (4)Multiple Access-FDMA, CDMA, TDMA, SDMA, DSSS, FHSS, ALOHA, PACKET RADIO...82 pages
- UNIK4230: Mobile Communications: Abul KaosherNo ratings yetUNIK4230: Mobile Communications: Abul Kaosher47 pages
- OAS053201 SPS V3 (DRA) Basic Data Configuration ISSUE 8.1No ratings yetOAS053201 SPS V3 (DRA) Basic Data Configuration ISSUE 8.1137 pages
- Multiple Access Techniques For Wireless Communications: School of Information Science and Engineering, SDUNo ratings yetMultiple Access Techniques For Wireless Communications: School of Information Science and Engineering, SDU79 pages
- Chapter 5 - 55multiple Access TechniquesNo ratings yetChapter 5 - 55multiple Access Techniques34 pages
- Wireless Communication Multiple Access Technique100% (2)Wireless Communication Multiple Access Technique76 pages
- Multiple Access Techniques For Wireless CommunicationsNo ratings yetMultiple Access Techniques For Wireless Communications92 pages
- 3.4 3.5 3.6 3.7 3.8 Multiple Access Techniques:: BroadcastingNo ratings yet3.4 3.5 3.6 3.7 3.8 Multiple Access Techniques:: Broadcasting13 pages
- Department of Electronics & Communication Engineering: Satish Kanapala, Assistant ProfessorNo ratings yetDepartment of Electronics & Communication Engineering: Satish Kanapala, Assistant Professor43 pages
- Multiple Acces Techniques For Wireless Communication: Sammar Zahra International Islamic University, IslamabadNo ratings yetMultiple Acces Techniques For Wireless Communication: Sammar Zahra International Islamic University, Islamabad25 pages
- Chapter 6 - Multiplexing and Multiple AccessNo ratings yetChapter 6 - Multiplexing and Multiple Access32 pages
- Unit III, SC-Multiple Access TechniquesNo ratings yetUnit III, SC-Multiple Access Techniques17 pages
- Unit-Iv Multiple Division Access TechniquesNo ratings yetUnit-Iv Multiple Division Access Techniques16 pages
- Multiple Acces Techniques For Wireless Communication: Sammar Zahra International Islamic University, IslamabadNo ratings yetMultiple Acces Techniques For Wireless Communication: Sammar Zahra International Islamic University, Islamabad25 pages
- Multiple Acces Techniques For Wireless Communication: Sammar Zahra International Islamic University, IslamabadNo ratings yetMultiple Acces Techniques For Wireless Communication: Sammar Zahra International Islamic University, Islamabad25 pages
- Multiple Access Techniques in Wireless Communication PPT Ayaz100% (2)Multiple Access Techniques in Wireless Communication PPT Ayaz28 pages
- Multiple-Access Techniques: These Methods Can Be Combined E.g.: SDMA/FDMA/TDMA in GSMNo ratings yetMultiple-Access Techniques: These Methods Can Be Combined E.g.: SDMA/FDMA/TDMA in GSM24 pages
- Lecture-7 Investigating Network TrafficNo ratings yetLecture-7 Investigating Network Traffic44 pages
- Controller. The MAC Sub-Layer Acts As An Interface Between The Logical Link Control (LLC)No ratings yetController. The MAC Sub-Layer Acts As An Interface Between The Logical Link Control (LLC)7 pages
- ##Multiple Access:: Duplexer Is Used Inside Each Subscriber Unit and Base Station To Allow Simultaneous RadioNo ratings yet##Multiple Access:: Duplexer Is Used Inside Each Subscriber Unit and Base Station To Allow Simultaneous Radio5 pages
- Pexip Infinity Hyper-V Installation Guide V34.aNo ratings yetPexip Infinity Hyper-V Installation Guide V34.a18 pages
- GSM Layer III Messages: Prepared By, Syed Masroor Ali PakistanNo ratings yetGSM Layer III Messages: Prepared By, Syed Masroor Ali Pakistan47 pages
- Introduction To Electronics Communication System100% (1)Introduction To Electronics Communication System8 pages
- TTL Dual Last Mile Solution For Zealous Services-Chennai Scenario-Active-ActiveNo ratings yetTTL Dual Last Mile Solution For Zealous Services-Chennai Scenario-Active-Active8 pages
- Computer Networks: Hub, Switch, Router, BridgeNo ratings yetComputer Networks: Hub, Switch, Router, Bridge34 pages
- Study & Simulation of Vehicular Ad-Hoc Networks Using MATLAB75% (8)Study & Simulation of Vehicular Ad-Hoc Networks Using MATLAB5 pages
- Design of An Intelligent Nitrogen Injection System For Fire Saftey of Power TransformerNo ratings yetDesign of An Intelligent Nitrogen Injection System For Fire Saftey of Power Transformer6 pages
- Cambium Networks Data Sheet Cnmatrix EX2000 Series SwitchesNo ratings yetCambium Networks Data Sheet Cnmatrix EX2000 Series Switches10 pages
- Investigating 802.11A Wireless Standard For High Speed Network Using Different Modulation TechniquesNo ratings yetInvestigating 802.11A Wireless Standard For High Speed Network Using Different Modulation Techniques5 pages
- Satellite Router Quick Start: Odin GuideNo ratings yetSatellite Router Quick Start: Odin Guide2 pages
