Chapter 4
Iterative Source–Channel Decoding1
Code concatenation constitutes a convenient technique for constructing powerful codes capable of achieving huge coding gains, while keeping the decoding complexity manageable. The concept of Serially Concatenated Codes (SCCs) having an ‘outer’ and an ‘inner’ code used in cascade was proposed in [125]. The discovery of Parallel Concatenated ‘turbo’ Codes (PCCs) [69] considerably improved the achievable performance gains by separating component codes through interleavers and using iterative decoding in order to further reduce the BER. The intensive research efforts of the ensuing era demonstrated [70] that the employment of iterative processing techniques is not limited to traditional concatenated coding schemes. In other words, the ‘turbo principle’ [53] is applicable to numerous other algorithms that can be found in digital communications, for example in turbo equalization [82], spectrally efficient modulation [83, 84], turbo multiuser detection [87, 88] and channel-coded code-division multiple-access [90].
Previously, substantial research efforts have been focussed on optimizing concatenated coding schemes in order to improve the asymptotic slopes of their error probability curves, especially at moderate to high SNRs. Recently, researchers have focussed their efforts on investigating the convergence behavior of iterative decoding. In [77] the authors proposed a so-called density evolution algorithm for calculating the convergence thresholds ...
