Serial Concatenated Codes Based on Iterative and Non-Iterative Decoding

Abstract

This paper gives the design and implementation of iterative and non-iterative decoders of serially concatenated block and convolutional coding schemes. The maximum likelihood (ML) decoder is implemented by applying Viterbi algorithm to trellis of the code. The performance of non-iterative decoder has been improved by using soft input soft output (SISO) ML-decoder as inner decoder and soft decision decoder as output decoder. Iterative SISO ML decoding of product codes is implemented using Pyndiah’s iterative decoder. The log-maximum a posteriori (Log-MAP) decoder is used in iterative decoding of serially concatenated codes (SCCs). To assess the performance of iterative and non-iterative decoders, simulation results for serially concatenated codes transmitted over AWGN channel, with low SNRs (power limited channel like deep space communication channel), are presented. The simulation process includes studying the bit error rate (BER) performances of serially concatenated codes with different parameters like, code dimension, minimum Hamming distance of the outer code, number of trellis sections (for block codes), number of memories, free distance, and encoder type (for convolutional codes). The influence of these parameters on interleaving gain and bit error rate of iterative and non-iterative decoders is discussed.