Convolutional decoders based on artificial neural networks

This paper investigates new methods of decoding convolutional codes based on neural networks. The methods are compared using BER curves obtained by simulation. New algorithms, based on iterative decoding, simulated annealing and total search, are investigated and the results obtained are presented. Both the neural network decoder and the Viterbi decoder are simulated and the bit error rates are compared. It is seen that the BER curves of the neural network decoders compare well with and even outperforms that of the decoder based on Viterbi algorithm. It was shown that the novel decoding algorithm based on total search gives the results that are comparable with or better than the results obtained by using turbo decoding techniques.