Stream-oriented turbo codes

A stream-oriented parallel concatenated convolutional (turbo) code (PCCC) is examined. The stream encoder uses continuous encoding with the the interleaver(s) viewed as a finite state permuters (FSPs). With this encoding architecture, block interleaver design is reexamined and convolutional interleavers are considered. Decoding is performed using a pipelined architecture where each constituent decoder uses a continuous version of the symbol-by-symbol MAP decoding algorithm. Performance and implementation issues are discussed and comparisons made to the traditional block decoding architecture. Finally, stream PCCC design using convolutional interleavers is examined based on performance and implementation