Wireless Standard Identification Based on Extended Radial Basis Function Neural Network in Cognitive Radio Het Nets
Monika Tulsyan1, Seemanti Saha2, Rajarshi Bhattacharya3

1Monika Tulsyan, Department of ECE, NIT Patna, Patna (Bihar), India.
2Seemanti Saha, Department of ECE, NIT Patna, Patna (Bihar), India.
3Rajarshi Bhattacharya, Department of ECE, NIT Patna, Patna (Bihar), India.

Manuscript received on 18 April 2019 | Revised Manuscript received on 25 April 2019 | Manuscript published on 30 April 2019 | PP: 1639-1644 | Volume-8 Issue-4, April 2019 | Retrieval Number: D6744048419/19©BEIESP
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)

Abstract: This paper presents a novel scheme for the automatic identification of primary user (PU) signal with respect to known wireless standards like GSM, Bluetooth, WLAN a/b/g/af, Zigbee, LTE etc. in a heterogeneous Cognitive Radio Network. The Secondary users (SUs), aware of the coexisted PU signal standard, can better exploit the available spectrum with coexisting PUs. Hence, in the proposed work, an Extended Radial Basis Function (ERBF) neural network (NN)) is used to classify PU signals of various standards using relevant explicit and implicit features (extracted from the detected PU signal), as the input to the classifier. The proposed method also involves automatic classification of digital modulation format used by the PU signal without a prior knowledge of the signal parameters. The proposed method can recognize single carrier modulation schemes like MPSK, MFSK, MASK, and MQAM, along with multicarrier modulation scheme like OFDM. The recognized modulation format is utilized further as an implicit feature given as an input to the signal classifier. Extensive simulation in MATLAB has been carried out for various signal-to-noise ratio (SNR) ranging from -20 dB to 20 dB. Simulation studies show that the proposed classification method is giving excellent detection and classification performance of 98% for most of the wireless standards, even in very low SNR of -20 dB.
Keywords: Cognitive HetNet, Feature Extraction, Modulation Classification, Radial Basis Function, Wireless Standard Classification.

Scope of the Article: Classification