Convolutional codes for finite state channels

Document Type

Conference Proceeding

Publication Date



Summary form only given, as follows. Error control strategies for finite-state channels often utilize decoders designed for use over memoryless channels (memoryless decoders). These are used with either specifically designed codes, such as burst-error-correcting codes, or with random-error-correcting codes in conjunction with interleaving. It may be possible, however, to improve the error probability performance of coding schemes by employing decoders which exploit the memory in the channel. Such decoders would utilize, directly or indirectly, the channel state sequence information obtained in the channel output sequence. The notion of state in convolutional codes and the existence of efficient decoding algorithms which are sequential (e.g., Viterbi algorithm) make convolutional codes suitable for adaptation for use with decoders which utilize the memory in the finite-state channel without greatly increasing the concomitant complexity. Several new decoders for the use of convolutional codes over such channels are proposed. The main feature of these decoders is that they can be implemented sequentially on a trellis by a Viterbi-like algorithm. The performance of these decoders is evaluated by simulation and compared with the performance of memoryless decoders with and without interleaving. The results indicate that in cases in which the channel is bursty, these decoders significantly outperform the memoryless decoders. The cost of this improved performance is a somewhat moderate increase in the decoders' complexity.

First Page


This document is currently not available here.