Multiple Access Schemes

1
Multiple Access Schemes

• CDMA
• Direct Sequence Spread Spectrum (DS-SS)

2
Multiple Access Schemes

• Why use Multiple Access?
• Allows an entry point to multiple users to share
  a common channel.
• We will briefly review the following Multiple
  Access Schemes
• Time Division Multiple Access (TDMA)
• Code Division Multiple Access (CDMA)
• Frequency Division Multiple Access (FDMA)
• Hybrid Systems

3
Multiple Access Techniques (TDMA)

• Advantages/Disadvantages of TDMA
• Uses only one carrier at a time
• No intermodulation impairment
• Achieves selectivity in time domain
• Selectivity is simpler than FDMA
• Ideal for digital communications
• Ideal for satellite on-board processing
• Utilizes maximum power capability and being BW
  efficient
• Overall is more complex and costly compared to
  FDMA

4
Multiple Access Techniques (TDMA) (cont)
5
Multiple Access Techniques (FDMA)

• Share a common transponder spectrum
• The whole spectrum is divided into subbands
• Advantages/Disadvantages of FDMA
• The overall FDMA method is simpler than TDMA
• Complicated bandpass filters are required
• Strict linearity requirement of the medium
• Normally should back off from saturation
• Cross-talk problem (not a problem in TDMA)
• Lack of timing requirement, attractive for fading
  channels

6
Multiple Access Techniques (FDMA) (cont)
7
Multiple Access Techniques (CDMA)

• Transmit multiple signals in the same frequency
  band at the same time
• Using orthogonality principle
• Orthogonal codes/spread spectrum
• Uses overlapping spectrum
• Typically used for military, data collection, and
  digital applications
• Frequency hopping, DS, FSK
• Synchronization is required
• For securityjam resistance

8
Multiple Access Techniques (CDMA) (cont)
9
Code Division Multiple Access (CDMA)

• All users have a Psuedo-Random Noise (PN)
  sequence code that is essentially uncorrelated
  with that assigned to any other user in the
  channel.

10
Code Division Multiple Access (CDMA) (cont)

• Direct Sequence This spread spectrum technique
  encodes each bit of information with many chips.
  Thus the frequency content of the signal has
  changed in order to take better advantage of the
  wireless channel.
• Consider a bit stream with Rb 10KBps and say
  each bit stream is coded with 100 chips, thus the
  chip rate becomes Rc 1MCps.

Spreading
11
Code Division Multiple Access (CDMA) (cont)

• Direct Sequence This spread spectrum technique
  is measured by what is called the Processing
  Gain (PG), which is defined as
• For the example shown earlier, we have

12
Direct Sequence Spread Spectrum (DS-SS)
spreading
Transmitter BPSK
Receiver
13
Code Generation in DS-SS
DS modulator
maximal length (ML) polar spreading code
14
A QPSK-DS Modulator
Constellation diagram
QPSK-modulator

• After serial-parallel conversion (S/P) data
  modulates the orthogonal carriers
• Modulation on orthogonal carriers spreaded by
  codes c1 and c2
• Spreading codes c1 and c2 may or may not be
  orthogonal (System performance is independent of
  their orthogonality, why?)
• What kind of circuit can make the demodulation
  (despreading)?

15
Maximal Length Codes (? sequences)
ML code generator

• code determined by feedback taps
• code rate determined by clock rate

Maximum-length codes. For every integer k 3
there exists a maximum length code (n,k) with n
2 k - 1, dmin 2 k-1.
16
Maximal Length Codes (cont.)
17
Design of Maximal Length Generators by a Table
Entry

• Have very good autocorrelation but cross
  correlation not granted
• Are linear, cyclic block codes - generated by
  feedbacked shift registers
• Number of available codes depends on the number
  of shift register stages
• Code generator design based on tables showing tap
  feedbacks

5 stages-gt6 codes, 10 stages -gt60 codes, 25
stages -gt1.3x106 codes
For the formula see Peterson, Ziemer
Introduction to Spread Spectrum Communication,
p. 121
18
Maximal Length Codes (cont.)

• Feedback connections can be written directly from
  the table

19
Autocorrelation properties

• G?a t?? a???????s? t?? spreading code
  ???s?µ?p?????ta? ?? p???e?? t?? a?t?s?s??t?s??
  ?a? t?? ete??s?s??t?s??, d??t? st? d??t? t? s?µa
  p?? ?aµß??eta? ?ata???? p???ap?as???eta?
  (a?t?s?s?et??eta?) µe t?? spreading code ???e
  user.
• ? s????t?s? t?? a?t?s?s??t?s?? ???s?µ?p??e?ta?
  ??a t? µ?t??s? t?? ??a??t?ta? d?????s?? µ?a?
  a???????a? ??d??a.
• Rxx (t)1/? ? ?(t) ?(tt) dt
• ? Rxx (k) 1/N S x(n) x(nk) d?a???t?? ??????
  s?µa
• ? s????t?s? t?? ete??s?s??t?s?? ?p??????e? t??
  ??a??t?ta d?????s?? µeta?? d?? d?af??et????
  a????????? ??d??a.
• Rxy (t)1/? ? ?(t) Y(tt) dt
• Rxy (k) 1/N S x(n) y(nk) d?a???t?? ??????
  s?µa

20
Autocorrelation properties of M sequences

• G?a t?? a???????a X 1,1,1,-1,-1,1,-1 ? t?µ?
  t?? a?t?s?s??t?s?? e??a?
• Rxx autocorr (X)
• ?p?t??esµa
• Rxx 7,-1,-1,-1,-1,-1,-1
• G?a t?? a???????e? X1,1,1,-1,-1,1,-1 ?a?
  Y1,-1,1,-1,1,-1,1 ? t?µ? t?? ete??s?s??t?s??
  e??a?
• Rxy crosscorr (X,Y)
• ?p?t??esµa
• Rxy -1,3,-1,3,-5,3,-1

21
Autocorrelation properties of M sequences

• ?? M-sequence pa?????ta? ap? ap?? LSR (linear
  shift registers).
• ? t?µ? a?t?s?s??t?s?? t??? d??eta? ap? t?? e???
  t?p?
• 1 t 0 mod Nc
• Rxx(t)
• -1/Nc d?af??et???
• ?? t?µ?? ete??s?s??t?s?? t?? a????????? p??pe? ?a
  e??a? -t(n), t(n)-2 ?p??
• 1 2 (n1)/2 nodd
• t(n)
• 1 2 (n2)/2 neven

22
Autocorrelation properties of M sequences (cont.)

• ?a??de??µa st? Matlab ????µe 3 registers ?a? ??
  a?????? t?µ?? t?? register e??a? 1,0,1 ?a? ??
  ??µß?? a?at??f?d?t?s?? e??a? ? 1?? ?a? ? 3??
• m1mseq(3, 1,3, 1,0,1)
• m2mseq(3, 2,3, 1,1,1)
• ?p?t??esµa
• m1(1 0 1 0 0 1 1)
• m2(1 1 1 0 0 1 0)
• (1-gt1, 0 -gt -1) Polar
• ?? t?µ?? a?t?s?s??t?s?? ?a? ete??s?s??t?s?? e??a?
  
• autocorr(m1) (7,-1,-1,-1,-1,-1,-1)
• autocorr(m2)( 7,-1,-1,-1,-1,-1,-1)
• crosscorr(m1,m2)(3,-1,-1,-5,3,3,-1)
• ?? t?µ?? a?t?s?s??t?s?? ?a? ete??s?s??t?s?? e??a?
  ?? ??t??µe?e? ?p?te
• a?t?? ?? ?-sequence e??a? p??t?µ?µe?? ?e?????.

23
Gold sequences

• ? Gold Sequence pa???eta? ap? t? ap???e?st??? OR
  (EXOR) d?? M-sequences ?? ?p??e? ap?te???? ??a
  p??t?µ?µe?? ?e?????.
• ?? t?µ?? t?? ete??s?s??t?s?? t?? a?????????
  p??pe? ?a e??a? -1, -t(n), t(n)-2 ?p??
• 1 2 (n1)/2 nodd
• t(n)
• 1 2 (n2)/2 neven
• G?a ?a ?????? ??p??a p??ß??µata p??
  pa???s?????ta? st?? Gold sequence p??st??eta?
  st?? a???????a ??a chip (st?? a??? ? st? t????).
  ? sequence p?? p????pte? ???µ??eta? Orthogonal
  Gold sequence.
• ? t?µ? t?? ete??s?s??t?s?? t?? ?rthogonal Gold
  sequence e??a? 0 st? s?µe?? s???????sµ??. Sta
  ???a s?µe?a ta ?a?a?t???st??? e??a? ta ?d?a µe
  a?t? t?? Gold sequence.

24
Gold sequence µe Matlab

• ?st? ?t? ????µe t? e??? p??t?µ?µe?? ?e?????
  M-sequence
• m1mseq(3, 1,3, 1,1,1)
• m2mseq(3, 2,3, 1,1,1)
• ? Gold sequence p?? ?a pa?a??e? e??a?
  g1goldseq(m1, m2, 2)
• ? a???µ?? t?? Gold Sequences p?? pa?????ta? ???e
  f??? e??a? 2n 1 ?a? ?aµß??eta? µe t?? a??a?? t??
  a?????? t?µ?? t?? register ?a? t?? e?sa???? d??
  M-sequence ?ta? ???s?µ?p????µe ??a? register
  n-ad???? µetat?p?s??.
• ?p?t??esµa g1 0 0 0 0 1 1 0
• 1 0 0 1 1 0 1
• ?? t?µ?? a?t?s?s??t?s?? ?a? ete??s?s??t?s??
  e??a?
• autocorr(g1(1,)) (7,3,-1,-1,-1,-1,3)
• autocorr(g1(2,)) (7,-1,-5,3,3,-5,-1)
• crosscorr(g1(1,),g1(2,)) (-1,3,-1,-5,-1,3,-1)
  
• ? Gold sequence e??a? de?t? ??at? pa???e? t??
  ??t??µe?e? t?µ?? a?t?s?s??t?s?? ?a?
  ete??s?s??t?s??.

25
BER in AWGN
26
BER in Rayleigh Fading
27
Code Division Multiple Access (CDMA) (cont)

• Frequency Hopping This spread spectrum technique
  uses N frequency channels to transmit the data
  signal. The N channels are hopped in a
  pre-determined, but random, manner.
• Assume we have N channels, with Rb 10KHz. The
  transmit bandwidth now becomes spread from 10KHz
  to 1MHz.

28
Code Division Multiple Access (CDMA) (cont)

• Frequency Hopping This spread spectrum technique
  is measured by what is called Processing Gain
  (PG) which is defined as
• There are two classifications that depend on the
  hopping rate, fh
• Fast Hopping We hop more than one channel for
  the duration of the symbol.
• Slow Hopping We hop one channel for the duration
  of many symbols.

29
Code Division Multiple Access (CDMA) (cont)

• Lets consider a simple BPSK CDMA system
• Transmitter Architecture
• Receiver Architecture

30
Code Division Multiple Access (CDMA) (cont)

• Direct Sequence Interference Reduction
  (suppression) is achieved as follows.

31
Code Division Multiple Access (CDMA) (cont)

• Multipath Performance
• We will show how CDMA handles multipath signals
  by the following two cases at hand.
• ? lt Tc For this case, the multipath received
  signals are unresolved because the delay spread
  causes inter-chip ISI.
• ? gt Tc Here we can resolve the multipath
  received signal. The delayed interfering signal
  is handled as interference and is attenuated by
  the processing gain. This arrives at the term
  Time Diversity. One particular chip now arrives
  at different times where its pulses are
  non-overlapping. A RAKE receiver is used here.

32
Code Division Multiple Access (CDMA) (cont)

• Psuedo Noise Sequence Generator (PN)
• The output chip streams are usually generated by
  Linear Maximal Length Sequence Generators. These
  sequences are easy to generate and, depending on
  the generator polynomial, have very interesting
  properties. We prefer the polynomials to be of
  the primitive class. Such polynomials give three
  properties
• Look Like Random Data There will be a balance of
  1s and 0s where the period is defined as (N No.
  of shift registers)
• Cyclic Addition If one were to cyclically shift
  the PN sequence by X shifts, it would produce
  another codeword that when added to a codeword of
  Y shifts, results in still a third codeword of Z.
• Good Autocorrelation Property You want to be
  able to confidently decide whether this is your
  codeword or not.

33
Code Division Multiple Access (CDMA) (cont)

• Lets consider a simple example of two users
• The Base Station looks like
• The transmitted signal is represented as

34
Code Division Multiple Access (CDMA) (cont)

• The receiver architectures are shown below.

35
Some CDMA Issues

• Chip rate increases, then bandwidth increases
  (efficiency), PG increases, then performance
  improves
• Acquisition time (coarse resolution)
• Tracking performance (coarse resolution)
• RAKE receiver (typically 3 fingers are used)
• Power control (Near-Far problem)

36
Multiple Access Techniques (SDMA)

• Channel separation is spatial
• Makes use of orthogonality in geometry
• Beams, orbits, or polarization
• Spectrum is reused
• Spectrum conservation best
• Satellite complexity