Evaluation of block turbo code performance with the reduced search trellis decoding method

Iterative decoding of parallel concatenated codes, also known as turbo codes, can provide tremendous coding gain approximating to the Shannon limit. Application of turbo coding techniques to block codes gives many advantages, due to the finite code length of the constituent codes and their ease of modification. The authors investigate the application of the trellis decoding method to block turbo codes, since one of the most easy and efficient ways of providing a soft output at the decoder is incorporation of the trellis decoding method. The soft output Viterbi algorithm is mainly considered, with performance improvement techniques. The authors propose efficient reduced search algorithms in which the decoder complexity can be reduced about ten times with less than 0.1 dB degradation in coding gain