Soft maximum likelihood decoding of codes with partially unit memory

The method of soft decoding with the use of maximum likelihood principle of partial unit memory codes and whole unit memory codes is proposed. The codes with partially unit memory are convolutional codes with the length of code restriction equal to one. The decoding algorithm has less complexity than Viterbi algorithm when the code rate is more than 0,5. Proposed method uses Viterbi algorithm with the change of transition from one level of the array to another one by means of application of Wolf's sindrom decoding method. The method can be used both for soft and hard decoding. The example is considered.