Machine Learning-Aided, Robust Wideband Spectrum Sensing for Cognitive Radios

Abstract

In this project, a compressive-sampling based, robust spectrum sensing approach was developed for wideband cognitive radios. The compressive-sampling based front-end is intended for overcoming the hardware imposed limitations on wideband spectrum sensing while robust detection principles are used to obtain a spectrum sensing approach that helps alleviate sensitivity to non-Gaussian noise and interference. Simulations were carried out to demonstrate that even with reduced number of samples, the proposed compressive-sampling based robust detector can indeed provide either comparable or better results to that observed with conventional periodogram, but with significantly higher number of samples. These results encourage further investigations and improvements of the proposed approach as a viable candidate for the front-end processing of a wideband cognitive radio.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jun 12, 2015
Accession Number
ADA625246

Entities

People

  • Sudharman Jayaweera

Organizations

  • University of New Mexico

Tags

Communities of Interest

  • Autonomy
  • Space

DTIC Thesaurus Topics

  • Air Force
  • Air Force Research Laboratories
  • Algorithms
  • Cognitive Radio
  • Compressed Sensing
  • Detection
  • Detectors
  • Frequency
  • Frequency Domain
  • Gaussian Noise
  • Government Procurement
  • Machine Learning
  • Radio Equipment
  • Sampling
  • Signal Processing
  • Spacecraft
  • Spectra

Readers

  • Computer Vision.
  • Energy Conservation and Renewable Energy Engineering.
  • Radar Systems Engineering.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference