Answers are grounded in this paper and cite their sources.
An optimal interleaving between two component encoders of a turbo‐code is proposed. The optimally criterion for any real constructable interleaver is given. For component codes with known weight distribution (WD), the WD of the turbo‐code with optimal interleaving is calculated. Furthermore, a method for approximate calculation of the WD of convolutional code with terminated feed‐forward and feed‐back encoders is proposed. Using this WD's and proposed interleaving, the union upper bounds on the bit error rate of the turbo‐code are given. It is shown that the often observed “break” in the performance curves of turbo‐codes is a result of their “broken” WD. For component codes lying on the Gilbert‐Varshamov bound, an upper bound on the ratio between the minimum distance of the turbo‐code and the component code is determined. The lower bounds on the signal‐to‐noise ratio needed to achieve error free decoding of turbo‐codes with random component coders are given. These bounds are compared with the Shannon limit and analogous bounds for random codes.
No takes yet. Share an insight, caveat, or question.
Y.V. Svirid (1995) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: