A method for recognition and classification for hybrid signals based on Deep Convolutional Neural Network

This paper presents a method for hybrid radar and communication signal recognition and classification based on Deep Convolutional Neural Network (DCNN) for feature extraction and classification. The main idea is to transform modulation mode into time-frequency map for recognition. To overcome the single input single output (SISO) characteristic of DCNN to output multi tags for hybrid signals, we propose a repeated selective strategy that segments the time-frequency map and classifies the selective regions by utilizing DCNN repeatedly. The experiments compare traditional methods (ANN and SVM, which can only satisfy the classification of single signals) with our method and show that the DCNN with short-time Fourier transform (STFT) performs better and more stable in single signals and achieve a classification accuracy over 92% at 0 dB and over 98% at 5 dB in hybrid signals.

[1]  Zhou Yue,et al.  Modulation classification based on spectrogram , 2012 .

[2]  Trevor Darrell,et al.  Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation , 2013, 2014 IEEE Conference on Computer Vision and Pattern Recognition.

[3]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[4]  Timothy J. O'Shea,et al.  An Introduction to Machine Learning Communications Systems , 2017, ArXiv.

[5]  Taghi M. Khoshgoftaar,et al.  A survey of transfer learning , 2016, Journal of Big Data.

[6]  Junde Song,et al.  Signal Classification Based on Spectral Correlation Analysis and SVM in Cognitive Radio , 2008, 22nd International Conference on Advanced Information Networking and Applications (aina 2008).

[7]  Li Jingchun,et al.  Mixed Recognition Algorithm for Signal Modulation Schemes by High-order Cumulants and Cyclic Spectrum , 2016 .

[8]  Michael S. Bernstein,et al.  ImageNet Large Scale Visual Recognition Challenge , 2014, International Journal of Computer Vision.

[9]  Qi Zhu,et al.  Combining clustering and SVM for automatic modulation classification , 2012, Int. J. Comput. Appl. Technol..

[10]  S. A. Alshebeili,et al.  An overview of feature-based methods for digital modulation classification , 2013, 2013 1st International Conference on Communications, Signal Processing, and their Applications (ICCSPA).

[11]  Ali Abdi,et al.  Survey of automatic modulation classification techniques: classical approaches and new trends , 2007, IET Commun..

[12]  Nitish Srivastava,et al.  Dropout: a simple way to prevent neural networks from overfitting , 2014, J. Mach. Learn. Res..

[13]  Honglak Lee,et al.  Understanding and Improving Convolutional Neural Networks via Concatenated Rectified Linear Units , 2016, ICML.

[14]  Ying Wu,et al.  Signal Classification Method Based on Support Vector Machine and High-Order Cumulants , 2010, Wirel. Sens. Netw..

[15]  Jun Zhang,et al.  Feature extraction and image retrieval based on AlexNet , 2016, International Conference on Digital Image Processing.

[16]  Marion Berbineau,et al.  Automatic Modulation Recognition Using Wavelet Transform and Neural Networks in Wireless Systems , 2010, 2009 9th International Conference on Intelligent Transport Systems Telecommunications, (ITST).

[17]  Long Zhang,et al.  A modulation classification based on SVM , 2016, 2016 15th International Conference on Optical Communications and Networks (ICOCN).

[18]  Jing Li Wang,et al.  Color image segmentation: advances and prospects , 2001, Pattern Recognit..