Abstract | ||
---|---|---|
This paper presents a new approach to the automatic modulation classification (AMC) of cochannel signals based on deep learning techniques using convolutional neural network (CNN). Conventional approaches to this problem use features from higher order statistics and cyclic statistics. Data with long length is required to achieve good feature estimation and high classification rate. However, data with long length may cause problems such as latency to practical operations. By applying deep learning techniques based on CNN, AMC can be conducted directly by exploring raw features itself. Meanwhile, oversampled data is reshaped to a two-dimensional data matrix in order to take advantages of the image processing capability of CNN. One main advantage of this method is that much shorter data length is required to achieve good classification rate as compared to conventional approaches. Simulation results are provided to demonstrate the effectiveness of the proposed methods. |
Year | DOI | Venue |
---|---|---|
2018 | 10.1109/ICDSP.2018.8631682 | DSL |
Keywords | Field | DocType |
Modulation,Convolution,Signal to noise ratio,Deep learning,Neurons,Feature extraction,Kernel | Kernel (linear algebra),Pattern recognition,Computer science,Convolutional neural network,Convolution,Higher-order statistics,Signal-to-noise ratio,Image processing,Feature extraction,Artificial intelligence,Deep learning | Conference |
ISSN | ISBN | Citations |
1546-1874 | 978-1-5386-6811-5 | 1 |
PageRank | References | Authors |
0.37 | 0 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Jiajun Sun | 1 | 1 | 2.06 |
Guohua Wang | 2 | 9 | 4.89 |
Zhiping Lin | 3 | 291 | 37.46 |
Sirajudeen Gulam Razul | 4 | 96 | 14.41 |
Xiaoping Lai | 5 | 240 | 25.14 |