It has long been known that one of the ways to achieve very good error-correcting capabilitites is by using very long codes. With turbo codes we are able to construct very long block codes based on parallel concatenation of small convolutional codes. The turbo codes are decoded by an iterated process. Although the decoding is not maximum likelihood, the results are still very good. The research aims at further understanding of the coding scheme, especially the convergence of the iterated decoding, and at general improvement of the system. Further, we are investigating the possibilities of hardware implementations of the decoders and alternative coding shemes inspired by the turbo codes.
|Period||01/09/93 → …|