
i
i
i
i
i
i
i
i
3.2 Digital Communication 57
16 State
Decoder
DEC
1
16 State
Decoder
DEC
2
Interleaving
y
k
x
k
y
1k
y
2k
Latency: L
1
Latency: L
2
Λ
1
(d
k
) Λ
1
(d
n
)
d
n-L2
Figure 3.6 The decoding of Turbo codes.
evidence it receives from the incoming prior probabilities with the parity information
provided by the code. After a number of iterations, the decoder converges to an
estimate of the transmitted codeword.
The decoder of Turbo codes is shown in Figure 3.6, which is made up of two
elementary decoders (DEC
1
and DEC
2
) in a serial concatenation scheme. The first
elementary decoder DEC
1
is associated with the lower rate R
1
encoder C
1
and yields
a weighted decision. For a discrete