A serial design of iterative belief propagation decoders for convolutional codes

The belief propagation (BP) decoding algorithm may be suitable for the decoding of convolutional self-orthogonal codes which were originally proposed for one-step threshold decoding. In this paper, a serial design of iterative BP decoder for convolutional self-orthogonal codes is presented. Using the algebraic structures of convolutional codes, the iterative BP decoder is designed as a serial concatenation of several one-step BP decoders. These one-step BP decoders are implemented using mainly the shift registers in a structure similar to that of type-H threshold decoders. The iterative BP decoder performs a nontrellis-based forward-only algorithm and has only an initial decoding delay, thus avoiding intermediate decoding delays that usually accompany BP or turbo decoding of data frames. As shown by simulation results, the use of weighing techniques has provided substantial improvements to the error performance of the iterative BP decoding at a cost of several multipliers in hardware implementation. The iterative BP decoder may be attractive to the practical applications in very high data rate areas.