Research Article
Automatic Modulation Classification Using Convolutional Recurrent Attention Network
@INPROCEEDINGS{10.1007/978-3-030-89814-4_54, author={Yuzheng Yang and Sai Huang and Shuo Chang and Hua Lu and Yuanyuan Yao}, title={Automatic Modulation Classification Using Convolutional Recurrent Attention Network}, proceedings={Mobile Multimedia Communications. 14th EAI International Conference, Mobimedia 2021, Virtual Event, July 23-25, 2021, Proceedings}, proceedings_a={MOBIMEDIA}, year={2021}, month={11}, keywords={Automatic modulation classification Attention block Adversarial attack Convolution block Long short term memory network}, doi={10.1007/978-3-030-89814-4_54} }- Yuzheng Yang
Sai Huang
Shuo Chang
Hua Lu
Yuanyuan Yao
Year: 2021
Automatic Modulation Classification Using Convolutional Recurrent Attention Network
MOBIMEDIA
Springer
DOI: 10.1007/978-3-030-89814-4_54
Abstract
The development of wireless communication technology is much faster than the pace of its security. The interference such as adversarial attack degrades the accuracy and efficiency of communication environment. Automatic modulation classification (AMC) is viewed as an effective method to discover and identify the modulation mode of wireless signal corrupted by noise and interference. This paper proposes a novel modulation classification framework using Convolutional Recurrent Attention Network (CraNET) which is mainly composed of the convolution block and long short term memory network (LSTM) based attention block. Convolution block extracts the signal features in the feature extraction module. In the weighting module, the LSTM based attention block selectively weights the extracted features to weaken the content that has no contribution to the performance improvement. Extensive simulation verifies that the CraNET based modulation classification method performs higher accuracy and superior robustness than that of other existing methods.
