Spectrum Sensing for Cognitive Radio using Deep Autoencoder Neural Network and SVM

Cognitive radio is one of solution to increase the efficient usage of frequency by spectrum sharing. One important compenent in the spectrum sharing is spectrum management system. The spectrum management system distributes information of white space. The system collects the information by performing spectrum monitoring. Spectrum monitoring is peformed by spectrum sensing. Spectrum sensing is challanging in multi radio environment because it has to discriminate several different radio system. In this paper, we proposed a spectrum sensing method based on deep auto encoder neural network and suport vector machine (SVM). Auto encoder neural network is used for automatic feature learning. The output of the feature learning then will be classified using support vector machine (SVM). SVM will classify input signal as primary user or secondary user. Using the proposed method, the sensing accuracy achieve the good performance up to 95 %.

[1]  Yi Zhang,et al.  Spectrum Monitoring for Radar Bands Using Deep Convolutional Neural Networks , 2017, GLOBECOM 2017 - 2017 IEEE Global Communications Conference.

[2]  Ramjee Prasad,et al.  An alternative implementation of a cyclostationary detector , 2012, The 15th International Symposium on Wireless Personal Multimedia Communications.

[3]  Mohsen Guizani,et al.  Deep Features Learning for Medical Image Analysis with Convolutional Autoencoder Neural Network , 2017 .

[4]  B. Ramkumar,et al.  Automatic modulation classification for cognitive radios using cyclic feature detection , 2009, IEEE Circuits and Systems Magazine.

[5]  Jeffrey H. Reed,et al.  Cyclostationary Approaches to Signal Detection and Classification in Cognitive Radio , 2007, 2007 2nd IEEE International Symposium on New Frontiers in Dynamic Spectrum Access Networks.

[6]  Serge Andrianov,et al.  Comparison of Regularization Methods for ImageNet Classification with Deep Convolutional Neural Networks , 2014 .

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

[8]  Luiz A. DaSilva,et al.  Context-aware cognitive radio using deep learning , 2017, 2017 IEEE International Symposium on Dynamic Spectrum Access Networks (DySPAN).

[9]  Ingrid Moerman,et al.  End-to-End Learning From Spectrum Data: A Deep Learning Approach for Wireless Signal Identification in Spectrum Monitoring Applications , 2017, IEEE Access.

[10]  Andriyan Bayu Suksmono,et al.  A Blind Spectrum Sensing for Cognitive Radio Based on Jarque-Bera Normality Test , 2016 .

[11]  Lindsay Cain,et al.  Convolutional neural networks for radar emitter classification , 2018, 2018 IEEE 8th Annual Computing and Communication Workshop and Conference (CCWC).

[12]  A. B. Suksmono,et al.  A HOS-Based Blind Spectrum Sensing in Noise Uncertainty , 2015 .

[13]  Radha Poovendran,et al.  Incentivizing crowdsourcing for radio environment mapping with statistical interpolation , 2015, 2015 IEEE International Symposium on Dynamic Spectrum Access Networks (DySPAN).

[14]  Ben Y. Zhao,et al.  Towards commoditized real-time spectrum monitoring , 2014, HotWireless@MobiCom.

[15]  Janne Riihijärvi,et al.  Enabling LTE in TVWS with radio environment maps: From an architecture design towards a system level prototype , 2014, Comput. Commun..

[16]  Edward F. Drocella.  3.5 GHz Exclusion Zone Analyses and Methodology , 2015 .