A parallel processing scheme for high-speed implementation of turbo codes

An important characteristic of turbo codes is the iterative nature of their decoder. One of the factors in determining the performance of turbo decoder is the number of iterations that is used for decoding each input sequence (frame). Using more iterations means decoding more sequences correctly. For example a decoder with 6 iterations per frame is not capable of decoding sequences that may be corrected in 7, 8 or more iterations. On the other hand, in practice, increasing the number of iterations increases the decoding delay, hence, reducing the decoder throughput. So, we have turbo decoders with limited capability and fixed number of iterations. This project proposes a parallel processing scheme to increase the performance and/or the throughput of turbo decoders.