A soft decision output convolutional decoder based on the application of neural networks

Abstract

The paper investigates BER characteristics of a new algorithm for decoding convolutional codes based on neural networks. The novelty of the algorithm is in its capability to generate soft output estimates of the message bits encoded. It is shown that the defined noise energy function, which is traditionally used for the soft decoding algorithm of convolutional codes, can be related to the well known log likelihood function. The coding gain is calculated using a developed simulator of a coding communication system that uses a systematic 1/2-rate convolutional code