Comparison of RF spectrum prediction methods for dynamic spectrum access

Jacob A. Kovarskiy, Anthony F. Martone, Kyle A. Gallagher, Kelly D. Sherbondy, Ram Mohan Narayanan

Research output: Chapter in Book/Report/Conference proceedingConference contribution

4 Citations (Scopus)

Abstract

Dynamic spectrum access (DSA) refers to the adaptive utilization of today's busy electromagnetic spectrum. Cognitive radio/radar technologies require DSA to intelligently transmit and receive information in changing environments. Predicting radio frequency (RF) activity reduces sensing time and energy consumption for identifying usable spectrum. Typical spectrum prediction methods involve modeling spectral statistics with Hidden Markov Models (HMM) or various neural network structures. HMMs describe the time-varying state probabilities of Markov processes as a dynamic Bayesian network. Neural Networks model biological brain neuron connections to perform a wide range of complex and often non-linear computations. This work compares HMM, Multilayer Perceptron (MLP), and Recurrent Neural Network (RNN) algorithms and their ability to perform RF channel state prediction. Monte Carlo simulations on both measured and simulated spectrum data evaluate the performance of these algorithms. Generalizing spectrum occupancy as an alternating renewal process allows Poisson random variables to generate simulated data while energy detection determines the occupancy state of measured RF spectrum data for testing. The results suggest that neural networks achieve better prediction accuracy and prove more adaptable to changing spectral statistics than HMMs given sufficient training data.

Original languageEnglish (US)
Title of host publicationRadar Sensor Technology XXI
EditorsArmin Doerry, Kenneth I. Ranney
PublisherSPIE
Volume10188
ISBN (Electronic)9781510608771
DOIs
StatePublished - Jan 1 2017
EventRadar Sensor Technology XXI 2017 - Anaheim, United States
Duration: Apr 10 2017Apr 12 2017

Other

OtherRadar Sensor Technology XXI 2017
CountryUnited States
CityAnaheim
Period4/10/174/12/17

Fingerprint

Dynamic Spectrum Access
Frequency Spectrum
radio frequencies
Hidden Markov models
Neural networks
Prediction
predictions
Statistics
Recurrent neural networks
Markov Model
Bayesian networks
Multilayer neural networks
Cognitive radio
Random variables
Markov processes
Neurons
Alternating Renewal Process
Neural Networks
Brain
Energy Detection

All Science Journal Classification (ASJC) codes

  • Electronic, Optical and Magnetic Materials
  • Condensed Matter Physics
  • Computer Science Applications
  • Applied Mathematics
  • Electrical and Electronic Engineering

Cite this

Kovarskiy, J. A., Martone, A. F., Gallagher, K. A., Sherbondy, K. D., & Narayanan, R. M. (2017). Comparison of RF spectrum prediction methods for dynamic spectrum access. In A. Doerry, & K. I. Ranney (Eds.), Radar Sensor Technology XXI (Vol. 10188). [1018819] SPIE. https://doi.org/10.1117/12.2262306
Kovarskiy, Jacob A. ; Martone, Anthony F. ; Gallagher, Kyle A. ; Sherbondy, Kelly D. ; Narayanan, Ram Mohan. / Comparison of RF spectrum prediction methods for dynamic spectrum access. Radar Sensor Technology XXI. editor / Armin Doerry ; Kenneth I. Ranney. Vol. 10188 SPIE, 2017.
@inproceedings{9dbbb5a297534b97b9c7efd730e3583e,
title = "Comparison of RF spectrum prediction methods for dynamic spectrum access",
abstract = "Dynamic spectrum access (DSA) refers to the adaptive utilization of today's busy electromagnetic spectrum. Cognitive radio/radar technologies require DSA to intelligently transmit and receive information in changing environments. Predicting radio frequency (RF) activity reduces sensing time and energy consumption for identifying usable spectrum. Typical spectrum prediction methods involve modeling spectral statistics with Hidden Markov Models (HMM) or various neural network structures. HMMs describe the time-varying state probabilities of Markov processes as a dynamic Bayesian network. Neural Networks model biological brain neuron connections to perform a wide range of complex and often non-linear computations. This work compares HMM, Multilayer Perceptron (MLP), and Recurrent Neural Network (RNN) algorithms and their ability to perform RF channel state prediction. Monte Carlo simulations on both measured and simulated spectrum data evaluate the performance of these algorithms. Generalizing spectrum occupancy as an alternating renewal process allows Poisson random variables to generate simulated data while energy detection determines the occupancy state of measured RF spectrum data for testing. The results suggest that neural networks achieve better prediction accuracy and prove more adaptable to changing spectral statistics than HMMs given sufficient training data.",
author = "Kovarskiy, {Jacob A.} and Martone, {Anthony F.} and Gallagher, {Kyle A.} and Sherbondy, {Kelly D.} and Narayanan, {Ram Mohan}",
year = "2017",
month = "1",
day = "1",
doi = "10.1117/12.2262306",
language = "English (US)",
volume = "10188",
editor = "Armin Doerry and Ranney, {Kenneth I.}",
booktitle = "Radar Sensor Technology XXI",
publisher = "SPIE",
address = "United States",

}

Kovarskiy, JA, Martone, AF, Gallagher, KA, Sherbondy, KD & Narayanan, RM 2017, Comparison of RF spectrum prediction methods for dynamic spectrum access. in A Doerry & KI Ranney (eds), Radar Sensor Technology XXI. vol. 10188, 1018819, SPIE, Radar Sensor Technology XXI 2017, Anaheim, United States, 4/10/17. https://doi.org/10.1117/12.2262306

Comparison of RF spectrum prediction methods for dynamic spectrum access. / Kovarskiy, Jacob A.; Martone, Anthony F.; Gallagher, Kyle A.; Sherbondy, Kelly D.; Narayanan, Ram Mohan.

Radar Sensor Technology XXI. ed. / Armin Doerry; Kenneth I. Ranney. Vol. 10188 SPIE, 2017. 1018819.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Comparison of RF spectrum prediction methods for dynamic spectrum access

AU - Kovarskiy, Jacob A.

AU - Martone, Anthony F.

AU - Gallagher, Kyle A.

AU - Sherbondy, Kelly D.

AU - Narayanan, Ram Mohan

PY - 2017/1/1

Y1 - 2017/1/1

N2 - Dynamic spectrum access (DSA) refers to the adaptive utilization of today's busy electromagnetic spectrum. Cognitive radio/radar technologies require DSA to intelligently transmit and receive information in changing environments. Predicting radio frequency (RF) activity reduces sensing time and energy consumption for identifying usable spectrum. Typical spectrum prediction methods involve modeling spectral statistics with Hidden Markov Models (HMM) or various neural network structures. HMMs describe the time-varying state probabilities of Markov processes as a dynamic Bayesian network. Neural Networks model biological brain neuron connections to perform a wide range of complex and often non-linear computations. This work compares HMM, Multilayer Perceptron (MLP), and Recurrent Neural Network (RNN) algorithms and their ability to perform RF channel state prediction. Monte Carlo simulations on both measured and simulated spectrum data evaluate the performance of these algorithms. Generalizing spectrum occupancy as an alternating renewal process allows Poisson random variables to generate simulated data while energy detection determines the occupancy state of measured RF spectrum data for testing. The results suggest that neural networks achieve better prediction accuracy and prove more adaptable to changing spectral statistics than HMMs given sufficient training data.

AB - Dynamic spectrum access (DSA) refers to the adaptive utilization of today's busy electromagnetic spectrum. Cognitive radio/radar technologies require DSA to intelligently transmit and receive information in changing environments. Predicting radio frequency (RF) activity reduces sensing time and energy consumption for identifying usable spectrum. Typical spectrum prediction methods involve modeling spectral statistics with Hidden Markov Models (HMM) or various neural network structures. HMMs describe the time-varying state probabilities of Markov processes as a dynamic Bayesian network. Neural Networks model biological brain neuron connections to perform a wide range of complex and often non-linear computations. This work compares HMM, Multilayer Perceptron (MLP), and Recurrent Neural Network (RNN) algorithms and their ability to perform RF channel state prediction. Monte Carlo simulations on both measured and simulated spectrum data evaluate the performance of these algorithms. Generalizing spectrum occupancy as an alternating renewal process allows Poisson random variables to generate simulated data while energy detection determines the occupancy state of measured RF spectrum data for testing. The results suggest that neural networks achieve better prediction accuracy and prove more adaptable to changing spectral statistics than HMMs given sufficient training data.

UR - http://www.scopus.com/inward/record.url?scp=85021935497&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85021935497&partnerID=8YFLogxK

U2 - 10.1117/12.2262306

DO - 10.1117/12.2262306

M3 - Conference contribution

AN - SCOPUS:85021935497

VL - 10188

BT - Radar Sensor Technology XXI

A2 - Doerry, Armin

A2 - Ranney, Kenneth I.

PB - SPIE

ER -

Kovarskiy JA, Martone AF, Gallagher KA, Sherbondy KD, Narayanan RM. Comparison of RF spectrum prediction methods for dynamic spectrum access. In Doerry A, Ranney KI, editors, Radar Sensor Technology XXI. Vol. 10188. SPIE. 2017. 1018819 https://doi.org/10.1117/12.2262306