Research Article
Automatic Modulation Classification Using Dense Memory Fusion Network
@INPROCEEDINGS{10.4108/eai.27-8-2020.2294984, author={Anping Li and Juanjuan Huang and Xu Yang and Xiaofei Zhang and Meiying Wei}, title={Automatic Modulation Classification Using Dense Memory Fusion Network}, proceedings={Proceedings of the 13th EAI International Conference on Mobile Multimedia Communications, Mobimedia 2020, 27-28 August 2020, Cyberspace}, publisher={EAI}, proceedings_a={MOBIMEDIA}, year={2020}, month={11}, keywords={cognitive radio automatic modulation classification deep learning dense memory fusion neural network}, doi={10.4108/eai.27-8-2020.2294984} }
- Anping Li
Juanjuan Huang
Xu Yang
Xiaofei Zhang
Meiying Wei
Year: 2020
Automatic Modulation Classification Using Dense Memory Fusion Network
MOBIMEDIA
EAI
DOI: 10.4108/eai.27-8-2020.2294984
Abstract
Automatic modulation classification (AMC), as a key technology of cognitive radio (CR), aims to identify the modulation format of the received signal. In this paper, we propose a novel dense memory fusion neural network(DMFN) based AMC method where grid constellation matrix (GCM) extracted from the received signals with low computational complexity are utilized as the input of DMFN. In DMFN, densnet with densely connected structures is designed to extract high representative feature of GCMs, the unit of long short-term memory (LSTM) and fully connected layer are used to make classification decisions. Extensive simulations demonstrate that DMFN yields significant performance gain and takes higher robustness comparing with other methods. In addition, DMFN based AMC scheme achieves 90$\%$ classification accuracy at 4dB when the symbol length is set as 512, which illustrates its outstanding performance.