Cryptography Reference
In-Depth Information
•
The result is the MI of output IS corresponding to the MI of input IA.
e. An example
The simulations were performed on a 16-state binary turbo code with rate
2/3, with a pseudo-random interleaving of 20,000 bits. The decoding algorithm
is the MAP algorithm. Figure 7.27(a) shows the relation between IS and IA as
a function of the signal to noise ratio of the Gaussian channel.
Figure 7.28 - EXIT charts for different
E
b
/N
0
in the case of binary turbo codes,
rate 2/3, pseudo-random interleaving of 20,000 bits, (a) 16-state and (b) 8-state MAP
decoding.
EXIT chart
The extrinsic information transfer characteristic is now known for a SISO de-
coder. In the case of iterative decoding, the output of decoder 1 becomes
the input of decoder 2 and vice versa. Curves IS1
=
f
(
I
A1
=
IS2
)
and
IS2
=
f
(
IA2
=
IS1
)
, identical to one symmetry if the SISO decoders are the
same, are placed on the same graph as shown in Figure 7.27(b). In the case
of a high enough signal to noise ratio (here 2 dB), the two curves do not have
any intersection outside the point of coordinates (1,1) which materializes the
knowledge of the received message. Starting from null mutual information, it
is then possible to follow the exchange of extrinsic information along the iter-
