What is it about?
Modulation classification is about identification of modulation format of a received signal. It is performed prior to demodulation stage and after signal detection. Besides its utility in electronic warfare domain, modern communication systems and technologies such as cognitive radios and spectrum sensing make effective use of modulation classification for efficient spectrum utilization.
Featured Image
Why is it important?
In the recent past deep learning has found its place at the heart of almost all the engineering problems. The quest for achieving high end AI applications has also encouraged the communication engineers to use the advancements in deep learning to their advantage. We provide an effective utilization of one such kind of deep learning methods i.e. sparse autoencoder based deep neural networks (SAE-DNN) to blindly recognize the modulation format amidst frequency selective channel conditions in the presence of Doppler shift phenomenon. The diversity of proposed low-complexity feature set alongwith the strength of SAE-DNN provide a rather robust solution with substantially reduced computational complexity.
Perspectives
I believe this article will help the relevant researchers in this field to build upon it and further refine and improve the findings. This article provides motivation to utilize deep learning for various problems involving estimation and classification in the field of communication systems.
Maqsood Hussain Shah
Nanjing University of Aeronautics and Astronautics
Read the Original
This page is a summary of: A Robust Approach for AMC in Frequency Selective Fading Scenarios using Unsupervised Sparse Autoencoder based Deep Neural Network (SAE-DNN), IET Communications, November 2018, the Institution of Engineering and Technology (the IET),
DOI: 10.1049/iet-com.2018.5688.
You can read the full text:
Resources
Contributors
The following have contributed to this page
