Decision-making analysis in cooperative environments for decentralized cognitive radio networks

Currently, there are saturated frequency bands that affect the quality of service for new users. Cognitive radio provides an alternative solution to this problem through dynamic spectrum access. However, the solutions proposed in the current literature are focused on a centralized network and do not allow demonstrating the behavior in a multi-user environment, much less the effect that cooperation between secondary users can have. This article establishes a decision-making model for the best spectral opportunity selection with a cooperative approach in decentralized cognitive radio networks and contrasts its results with three multi-criteria decision-making algorithms: SAW, TOPSIS, and VIKOR. So, this research suggests a cooperative decision-making model based on four main modules. (1) a collaborative module for the exchange of information between SU; (2) a module for PU characterization; (3) a module of the probability of SU arrival; and (4) the SO feedback selection module. The results are obtained through simulations fed with experimental spectral occupancy data captured in a measurement campaign. Handoff and throughput were used as evaluation metrics, along with five levels of collaboration: 10%, 20%, 50%, 80%, and 100%, and eight different scenarios based on the type of network: GSM and Wi-Fi, the application type: real-time and best-effort, and the level of traffic: high and low. The contribution of this study lies in the fact that no current work includes the following relevant aspects for an adequate validation and evaluation of this proposal: First, the consideration of a decentralized cognitive radio network; second, the decision-making with cooperative strategies; third, different techniques for SO selection; fourth, the validation and evaluation with experimental spectral occupancy data captured in measurement campaigns; finally, the performance analysis in diverse networks, traffic levels, and types of applications.