Recursive Systematic Convolutional Research Articles

Fast Chase algorithm with an application in turbo decoding

Turbo product codes (TPCs) provide an attractive alternative to recursive systematic convolutional (RSC)-based turbo systems. Rather than employ trellis-based decoders, an algebraic decoder may be repeatedly employed in a low-complexity, soft-input/soft-output errors-and-erasures decoder such as the Chase algorithm. Taking motivation from efficient forced erasure decoders, this implementation re-orders the Chase algorithm's repeated decodings such that the inherent computational redundancy is greatly reduced without degrading performance. The result is a highly efficient fast Chase implementation. The algorithm presented here is principally applicable to single error-correcting codes although consideration is also given to the more general case. The new decoder's value in practical turbo schemes is demonstrated via application to decoding of the (64,57,4) extended Hamming TPC.

Simple hybrid type II ARQ technique using softoutputinformation

The authors present a simple automatic repeat request (ARQ) scheme for turbo codes which uses the soft output information, available as a by-product of the decoding process. To indicate the need for retransmission of a particular data block. The technique utilises the punctured data, not normally sent in recursive systematic convolutional (RSC) turbo code systems, to improve the operation of the decoder rather than the retransmission of the whole block, thus maximising the throughput. Furthermore, as no cyclic redundancy check (CRC) is included there is no overhead on transmission rate. The scheme is capable of correctly identifying a large proportion of all errors across a range of E/sub b//N/sub o/ whilst minimising incorrect retransmission requests.