An improved SOVA algorithm and DSP implementation for turbo codes decoding

SOVA has been a practical decoding algorithm for turbo-codes due to its low complexity and low latency of decoding. However, inaccuracy of the soft-decision output could lower the performance of SOVA. In this paper, we propose a modified SOVA by incorporating a corrective function to improve the accuracy of the soft decision output. The simulation results show that the modified corrective SOVA algorithm achieves significant performance improvement without an increase of decoding complexity. Implementation of the improved algorithm on a DSP is investigated by examining the effects of the degree of quantization and the finite word length on the performance and speed, as well as decoding memory storage requirements.