An End-to-End Demodulation System Based on Convolutional Neural Networks

Abstract

The demodulation of digital signal plays a key role in the communication system. The traditional demodulator is usually realized by special hardware platform, which has the disadvantages of high cost and long development cycle. In this paper, we propose an end-to-end digital signal demodulator based on convolutional neural network (CNN). It consists of an encoder and a decoder, in which the encoder encodes the input symbol sequence and maps the signal features to the hidden layer space. Then, the decoder decodes the features of the hidden layer space to obtain the demodulation result of the input sequence. The proposed algorithm can automatically learn how to demodulate the received signal without manually extracting the features. Compared with the traditional demodulator, the proposed CNN demodulator has better bit error rate (BER) performance.