Truncated Turbo Decoding with Reduced Computational Complexity

Abstract

The concept of concatenated codes and turbo decoding is well known and leads to a remarkably good performance in many applications. The resulting signal processing for this concept shows high complexity relative to conventional Viterbi decoding. This paper, therefore, considers an alternative concept of turbo decoding to reduce the computational complexity. In this case, those sections of the sequence to be decoded, where changes of bit decisions (compared to the previous iteration step) are very unlikely, are excluded from the soft-output viterbi algorithm (SOVA). This decoding is much easier to process and the loss of bit error rate (BER) performance is quite small or even negligible in comparison to conventional turbo decoding.