Recent applications of machine learning to modulation recognition have demonstrated the potential of deep learning to achieve state-of-the-art performance. We propose to further extend this approach by using flexible time-space decompositions that are more in line with the actual learning task, as well as integrate side-information, such as higher order moments, directly into the training process. Our promising preliminary results suggest that there are many more benefits to be reaped from such approaches.
@inproceedings{Arumugam2017a,
author = {Arumugam, Keerthi Suria Kumar and Kadampot, I. A. and Tahmasbi, M. and Shah, S. and Bloch, M. and Pokutta, S.},
booktitle = {Proc. IEEE Int. Symp. Dynamic Spectrum Access Networks (DySPAN)},
title = {Modulation recognition using side information and hybrid learning},
year = {2017},
address = {Piscataway, NJ},
month = mar,
pages = {1--2},
doi = {10.1109/DySPAN.2017.7920750},
file = {:2017-Arumugam-Dyspan.pdf:PDF},
groups = {Steganography and covert communications},
keywords = {learning (artificial intelligence), modulation, telecommunication computing, deep learning, flexible time-space decompositions, higher order moments, machine learning, modulation recognition, side information, Cognitive radio, Convolution, Dynamic spectrum access, Machine learning, Modulation, Network architecture, Signal to noise ratio}
}