Please use this identifier to cite or link to this item:
Title: Classification and reconstruction of communication signals based on convolutional neural network
Authors: Cai, Zhenmin
Keywords: DRNTU::Engineering::Electrical and electronic engineering
Issue Date: 2019
Abstract: Convolutional neural network (CNN) is now widely used in many areas including pattern recognition, intelligent control and computer science. CNN reduces the calculation of the model effectively and improves the robustness compared with Artificial Neural Network (ANN). This report uses a CNN model to do classification and features extraction on different modulation signals in communication. Besides that, robust signal reconstruction against noise is investigated based on the dictionary constructed using the features extracted by CNN. Firstly, a series of experiments to classify different kinds of modulation signals using CNN were done to verify the effectiveness CNN model in automatic feature extraction. One experiment was conducted on QAM and PSK signals to achieve an average signal classification accuracy of over 99%. Another experiment was done to classify 8PSK and QPSK and correct rate of this system reached 95% as well. Last but not least, we used convolutional sparse coding to reconstruct signals with the dictionary learnt by CNN model. The experiment shows that the dictionary is able to reconstruct signals even with low Signal Noise Ratio (SNR) and the atoms in the dictionary learned by CNN show different characteristics of signals.
Schools: School of Electrical and Electronic Engineering 
Fulltext Permission: restricted
Fulltext Availability: With Fulltext
Appears in Collections:EEE Theses

Files in This Item:
File Description SizeFormat 
  Restricted Access
3.63 MBAdobe PDFView/Open

Page view(s)

Updated on Jun 23, 2024


Updated on Jun 23, 2024

Google ScholarTM


Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.