Maximization of Throughput in Cognitive Radio using non-persistent CSMA combined with Reinforcement Learning

Abstract: The Cognitive Radio project aims to develop radios that can sense the existing spectrum and identify and use free frequency bands. Primary users using time division multiple access (TDMA) are licensed users who are assigned with certain channels, and secondary users using carrier sense multiple access (CSMA) are unlicensed users who are allowed to use the channels assigned to a primary user when it is free. This paper introduces a novel spectrum sharing protocol for secondary users combining multichannel non-persistent carrier sense multiple accesses (CSMA) and reinforcement learning. Simulation results show that the throughput performance improves when reinforcement learning is applied in conjunction with non-persistent CSMA.

Keywords: Cognitive Radio; multichannel CSMA; channel assignment; Reinforcement learning.