On the use of SOVA for iterative decoding

Most of the error correcting capabilities rely on the concept of iterative decoding. However, the iterative decoding needs a "soft-in/soft-out" component decoders. Such decoders can be realized either by using maximum a posteriori (MAP) symbol by symbol estimator or the soft output Viterbi algorithm (SOVA). The latter is more suitable for hardware implementation. Therefore, improving the performance of the SOVA is very important to future applications where novel class of codes such as turbo codes are used. In this paper, motivated by improving the hardware implementation, we study the relation between the updating depth, a parameter of the SOVA, and the number of iterations defined by the iterative decoding procedure.