Machine learning supervised antenna for software defined cognitive radios. Issue 9 (2nd September 2022)

• Record Type: Journal Article
• Authors: Goswami, Pankaj Kumar; Goswami, Garima
• Is Part Of: International journal of electronics. Volume 109:Issue 9(2022)
• Page Start: 1508
• Page End: 1531
• Publication Date: 2022-09-02
• Subjects: ANN -- CRN -- ML -- SDR -- WSN; Electronics -- Periodicals; Electronic control -- Periodicals; 621.381
• DOI: 10.1080/00207217.2021.1969447
• Languages: English
• ISSNs: 0020-7217

Abstract

ABSTRACT: Software-defined radio (SDR) deals with the opportunistic detection and allocation of the radio frequency (RF) signals in the cognitive radio network (CRN). The opportunistic spectrum use between primary licenced and secondary unlicensed user depends upon the adequate detection of the spectrum hole. This paper presents a design of a smart compact wideband patch antenna for SDR in cognitive radios. The fractal slot antenna with electromagnetic bandgap structure is designed on 12 × 18 mm 2 FR4 substrate as a primary perception module in the cognitive radio sensor network. The wide operating range of antenna 0.68–12.1 GHz is highly compatible with adaptive learning of SDR to deal with dynamic spectrum variation over a large bandwidth. The antenna exhibits high radiation efficiency with consideration of tangential losses of FR4 and sustainable gain 2.5 dBi over the complete impedance bandwidth. The proposed five-layered ANN-ML supervised SDR comprises multispectral resolution via coefficient estimation for primary and secondary user and elicits high throughput of proposed antenna during white space spectrum sensing. The data analytics for ANN-ML has modelled on Python 3.0 IDE and the results are validated to justify the candidature of the antenna for wideband sensing as per the Federal Communications Commission standards for CRN.