An improved design and SOVA algorithm for serial concatenated convolutional code

Serial concatenated code (SCC) has attracted lots of attention by its excellent performance when employing an iterative decoder. However, the previous research shows that it can obtain good performance only with excessive processing delay. It is analyzed that the decoding structure is the key point of its performance's convergence, as well as the decoding algorithm. This paper introduces the structure of the Log-MAP algorithm into the conventional SOVA decoder. The contribution of Log-MAP and SOVA avoids the updating of the matrices of the maximum path, and consequently makes a contribution to the requirement of short delay. Serial concatenated convolutional codes (SCCC) with different structures are also analyzed. The simulations of the improved algorithm for SCCC with short-frame interleaver prove that it is more suited to high-bit-rate low-delay communications systems.