Cryptography Reference
In-Depth Information
we have:
L ( c 3 )=ln 1+exp( L ( c 2 )+ L ( c 1 ))
exp( L ( c 2 )) + exp( L ( c 1 ))
L ( c 1 )
L ( c 2 )
(9.3)
Equation (9.3) enables us to define the switching operator
between the two
LLRs of the variables c 1 and c 2 .
Applying function tanh ( x/ 2) =
exp( x ) 1
exp( x )+1
to Equation (9.3), the latter be-
comes:
tanh L ( c 3 )
2
=
exp( L ( c 0 ))
1
exp( L ( c 0 ))+1 ×
exp( L ( c 1 ))
1
exp( L ( c 1 ))+1
j =0 tanh L ( c j )
(9.4)
1
=
2
It is practical (and frequent) to separate the processing of the sign and the
magnitude in Equation (9.4) which can then be replaced by the following two
equations:
1
sgn ( L ( c 3 )) =
sgn ( L ( c j ))
(9.5)
j =0
tanh |
=
tanh |
1
L ( c 3 )
|
L ( c j )
|
(9.6)
2
2
j =0
where the sign function sgn ( x ) is such that sgn ( x )=+1 if x
0 and sgn ( x )=
1
otherwise.
Processing the magnitude given by Equation (9.6) can be simplified by taking
the inverse of the logarithm of each of the terms of the equation, which gives:
= f 1
j =1 , 2
|
L ( c 3 )
|
f (
|
L ( c j )
|
)
(9.7)
where function f, satisfying f 1 ( x )= f ( x ) , is defined by:
f ( x )=ln(tanh( x/ 2))
(9.8)
These different aspects of the computation of function
will be developed in
the architecture part of this chapter.
Expression (9.6) in fact corresponds to the computation of (9.3) in the Fourier
domain. Finally, there is also a third writing of the LLR of variable c 2 [9.65, 9.23]:
L ( c 3 )= sign ( L ( c 1 )) sign ( L ( c 2 )) min (
|
L ( c 1 )
|
,
|
L ( c 2 )
|
)
ln (1 + exp (
−|
L ( c 1 )
L ( c 2 )
|
))
(9.9)
+ln(1+exp(
−|
L ( c 1 )+ L ( c 2 )
|
))
Search WWH ::




Custom Search