Joint Clustering Method and Reinforcement Learning-Based Spectrum Management in Mobile Cognitive Radio Networks

Abstract

The spectrum management scheme is a crucial issue to utilize the limited resource and secure a higher quality of performance in mobile cognitive radio networks. We propose a spectrum management method to utilize the spectrum effectively and minimize the communication cost in mobile cognitive radio networks. We consider the following parts such as statistical information about spectrum access patterns and the number of spectrum hand-off. In addition, this paper proposes to make the parameters to decide the spectrum access of the secondary user during the primary user doesn’t use it. Those parameters are based on dynamic clustering and reinforcement learning. However, both the dynamic clustering method and reinforcement learning haven’t been studied much to minimize communication costs in mobile cognitive radio networks. Firstly, our study is to improve the opportunity of spectrum access through the analysis based on the spectrum patterns and the number of both the secondary user and the primary user in mobile cognitive radio networks. Secondly, we decrease the number of hand-offs because it checks the status of the available channel and the patterns of the spectrum usage in the clustering. Above all, the improvement of this paper is to cut down the cost of communication by utilizing the way of the optimal spectrum access in mobile cognitive radio networks.