Simulation of a Cognitive Radio using Energy Detection & Reinforcement Learning

Brandon Dominique (New Jersey Institute of Technology)

In this paper, the concept, simulation and implementation of a Cognitive Radio (CR) is discussed. Because of the limited amount of space in the spectrum there needs to be a system in place that can efficiently allocate users to open parts of the spectrum, so that electronic messages of all types can be relayed successfully.  As autonomous vehicles (AV) increase in usage, the spectrum will become more crowded due to their different communication methods (i.e. vehicle to vehicle, vehicle to infrastructure). To resolve this, CRs can aid in the allocation of spectrum for AVs by detecting the presence of Primary Users (i.e. having priority access to a channel) on a given set of channels and avoiding any interference by selecting unused channels. Simulation of a CR is accomplished in this paper using Energy Detection and Reinforcement Learning algorithms such as Q-Learning and Upper Confidence Bound.

CAT Vehicle 2019

Brandon Dominique (New Jersey Institute of Technology)

Daniel Fishbein (Missouri State University)
 Kreienkamp (University of Notre Dame)

Alex Day (Clarion University of Pennsylvania)
Sam Hum (Colorado College)
Riley Wagner
 (the University of Arizona)

Eric Av (Gonzaga University)
Hoang Huynh (Georgia State University-Perimeter College)
John Nguyen (University of Minnesota, Twin Cities)

Brandon Dominique's experience: A brief overview of the work that I did at the University of Arizona for their student-led self driving car project, the CAT Vehicle.