A RNN Decoder for Channel Decoding under Correlated Noise

Abstract

We present and use a recurrent neural network as RNN decoder for the convolutional code under correlated noise. Using the bidirectional GRU(Gated Recurrent Unit) network and the fully connected neural network(DNN), the sequence feature information of the convolutional code is extracted, and the decoding result is calculated by the fully connected neural network. The RNN neural network decoder achieves better BER(Bit Error Rate) performance than the traditional Viterbi decoding algorithm for convolutional codes with shorter memory lengths, such as (2, 1, 3) convolutional codes under correlated noise. The greater noise correlation, the greater performance improvement of the decoder relative to the traditional Viterbi decoding algorithm. Besides, the RNN decoder is close to the MAP performance under the AWGN channel. In addition, the decoder is robust under different noise correlation models. Due to the limitation of the structure and complexity of the RNN decoder, as the memory length of the convolutional code increases, its decoding performance is gradually reduced, which is not suitable for convolutional codes with long memory lengths.