Deep Learning Based Automatic Modulation Classification With Long-Short Term Memory Networks

KARAHAN S. N., KALAYCIOĞLU A.

28th Signal Processing and Communications Applications Conference (SIU), ELECTR NETWORK, 5 - 07 October 2020, (Full Text) identifier identifier

  • Publication Type: Conference Paper / Full Text
  • Doi Number: 10.1109/siu49456.2020.9302280
  • Country: ELECTR NETWORK
  • Keywords: deep learning, recurrent neural networks, modulation classification, Bi-LSTM
  • Ankara University Affiliated: Yes

Abstract

The automatic modulation classification (AMC) process is used to determine the modulation format of the transmitted signal at the receiver side without any prior knowledge. Deep learning is a type of machine learning that consists of multiple layers in which raw data is taken as input. This study analyzes the AMC process with a deep learning approach. In this context, performances of LSTM (Long-Short Term Memory) and Bi-LSTM (Bidirectional LSTM) methods on the modulation classification problem are compared. Simulation results show that Bi-LSTM method has a higher performance than does the LSTM method.