ChiMaS: A spectrum sensing-based channels classification system for cognitive radio networks

Cognitive radio devices are able to sense the spectrum of frequencies and share access to vacant channels. These devices usually have a candidate channels list that must be sensed to find a vacant channel. In this paper, we propose a novel system called ChiMaS, which is able to manage the candidate channels list implementing three tasks: Analysis, Creation, and Sort. Analysis applies reinforcement learning algorithms to evaluate the channels quality based on their historical occupancy and their conditions; Creation is responsible for creating the Candidate Channels List; and Sort ranks the channels to obtain an Ordered Channels List in terms of quality. Results show that ChiMaS manages the candidate channels list following the IEEE 802.22 definition, while it finds the best channel in terms of availability and quality faster than Q-Noise+ algorithm, which was implemented for comparison purpose.