A Uniform Framework For Network Selection In Cognitive Radio Networks
Ye Wang,Jia Yu,Xiaodong Lin,Qinyu Zhang
DOI: https://doi.org/10.1109/ICC.2015.7248901
2015-01-01
Abstract:With the development of secondary spectrum markets, it is anticipated that multiple Primary Networks (PRNs) who own underutilized spectrum resources will be incorporated into Cognitive Radio Networks (CRNs). In this scenario, CRNs will have a greatly enhanced choice of accessible spectrum resources to support large volumes of Secondary Users (SUs), and guarantee the QoS reliability. Network selection problem, i.e. choosing which PRN to access, is essential for CRNs in a multi-PRN environment. However, to the best of our knowledge, there is still lack of a unified method to address the network selection problem. In this paper, we aim to present a uniform framework to investigate and evaluate network selection strategies for CRNs. First, we model the interactive process of SUs and PUs as a Continuous Time Markov Decision Process (CTMDP), and abstract the network selection strategy into the set of decision variables with respect to system states in the CTMDP. Second, under the proposed framework, we discuss multiple existing strategies, such as random, greedy, and statistically-weighted. Third, to achieve a more effective method, we derive the performance gradient of CRNs' utility function with respect to the network selection strategy, and propose a gradient-based optimal network selection strategy by using the theory of Markov performance potential. At last, simulations are conducted to validate the correctness of the proposed analytical framework, and the effectiveness of the proposed network selection scheme.