Low Complexity Autoencoder based End-to-End Learning of Coded Communications Systems

End-to-end learning of a communications system using the deep learning-based autoencoder concept has drawn interest in recent research due to its simplicity, flexibility and its potential of adapting to complex channel models and practical system imperfections. In this paper, we have compared the bit error rate (BER) performance of autoencoder based systems and conventional channel coded systems with convolutional coding (CC), in order to understand the potential of deep learning-based systems as alternatives to conventional systems. From the simulations, autoencoder implementation was observed to have a better BER in 0—5 dB E b /N 0 range than its equivalent half-rate convolutional coded BPSK with hard decision decoding, and to have only less than 1 dB gap at a BER of 10 -5. Furthermore, we have also proposed a novel low complexity autoencoder architecture to implement end-to-end learning of coded systems in which we have shown better BER performance than the baseline implementation. The newly proposed low complexity autoencoder was capable of achieving a better BER performance than half-rate 16-QAM with hard decision decoding over the full 0—10 dB E b /N 0 range and a better BER performance than the soft decision decoding in 0—4 dB E b /N 0 range.

Authors:
Rajapaksha Nuwanthika, Rajatheva Nandana, Latva-aho Matti

Publication type:
A4 Article in conference proceedings

Place of publication:
91st IEEE Vehicular Technology Conference, VTC Spring 2020

Keywords:
Autoencoder, channel coding, Deep learning, end-to-end learning, modulation, neural networks, wireless communications

Published:

Full citation:
N. Rajapaksha, N. Rajatheva and M. Latva-aho, “Low Complexity Autoencoder based End-to-End Learning of Coded Communications Systems,” 2020 IEEE 91st Vehicular Technology Conference (VTC2020-Spring), Antwerp, Belgium, 2020, pp. 1-7, doi: 10.1109/VTC2020-Spring48590.2020.9128456

DOI:
https://doi.org/10.1109/VTC2020-Spring48590.2020.9128456

Read the publication here:
http://urn.fi/urn:nbn:fi-fe202102185276