Title :
Correlation decoding of the (24, 12) Golay code using neural networks
Author :
Yuan, Jing ; Chen, C.S.
Author_Institution :
Inst. of Robotics & Intelligent Syst., Victoria Univ., BC, Canada
Abstract :
The (24, 12) extended Golay code contains eight disjoint subcodes. Each subcode can be constructed from 24 (real) orthogonal base vectors. It further contains 64 sub-subcodes that are equivalent to a three-dimensional universe code. This structure enables the authors to use partial correlation between the received vector and the base vectors to perform ideal correlation decoding. The correlation decoder can be implemented with a 64-flop and an 8-flop Hopfield neural network plus some housekeeping logic.<>
Keywords :
correlation theory; decoding; error correction codes; neural nets; (24, 12) extended Golay code; Hopfield neural network; correlation decoder; correlation decoding; disjoint subcodes; error correction codes; logic; orthogonal base vectors; partial correlation; received vector; sub-subcodes; three-dimensional universe code;
Journal_Title :
Communications, Speech and Vision, IEE Proceedings I
