Scribd
Upload
19 views20 pages

Kasami

The document discusses properties of binary random sequences and maximal length sequences (m-sequences) which are a type of pseudorandom sequence. M-sequences are generated deterministically using a linear feedback shift register with feedback taps corresponding to a primitive polynomial. M-sequences approximate random sequences and have desirable properties like good autocorrelation and spectral properties.

Uploaded by

Geleta Aman
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
19 views20 pages

Kasami

The document discusses properties of binary random sequences and maximal length sequences (m-sequences) which are a type of pseudorandom sequence. M-sequences are generated deterministically using a linear feedback shift register with feedback taps corresponding to a primitive polynomial. M-sequences approximate random sequences and have desirable properties like good autocorrelation and spectral properties.

Uploaded by

Geleta Aman
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 20

ECS455: Chapter 4

Multiple Access
4.5 m-sequence

Office Hours:
BKD 3601-7
Tuesday 9:30-10:30
Dr.Prapun Suksompong Tuesday 13:30-14:30
1 prapun.com/ecs455 Thursday 13:30-14:30
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.
 Some 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  1x
2 3

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  1x 3
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).
 Technically, 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.
 Therefore, we 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

You might also like