Theoretical interpretation and investigation of a 2/n rate convolutional decoder based on recurrent neural networks

In this paper a mathematical model of a 2/n rate conventional convolutional encoder/decoder system was developed to be applied for decoding using neural networks based on the gradient descent algorithm. The general expression for the energy function, needed for the recurrent neural networks decoding, is derived. Then, the expressions for the gradient decent updating rule are derived and the neural network decoder was designed. The coding scheme of a 2/3 rate code is simulated and results are compared with the results achieved in literature for one-input encoders.