The (d,k) subcode of a linear block code |
| |
Authors: | Patapoutian A Kumar PV |
| |
Affiliation: | IBM, Tucson, AZ; |
| |
Abstract: | A simple technique employing linear block codes to construct (d,k) error-correcting block codes is considered. This scheme allows asymptotically reliable transmission at rate R over a BSC channel with capacity CBSC provided R ⩽Cd,k-(1+CBSC), where Cd,k is the maximum entropy of a (d,k ) source. For the same error-correcting capability, the loss in code rate incurred by a multiple-error correcting (d,k) code resulting from this scheme is no greater than that incurred by the parent linear block code. The single-error correcting code is asymptotically optimal. A modification allows the correction of single bit-shaft errors as well. Decoding can be accomplished using off-the-shelf decoders. A systematic (but suboptimal) encoding scheme and detailed case studies are provided |
| |
Keywords: | |
|
|