Traffic pattern-based reward maximization for secondary user in dynamic spectrum access

In cognitive radio networks where secondary users (SUs) access spectral white spaces left by primary users (PUs), there exists a trade-off between sensing and transmission. However, the white spaces depend on the PUs' traffic patterns, and this fact can be utilized to optimize SU transmissions after traffic identification. This paper studies the effective approaches to adapt SU transmissions to the inherit dynamics associated to each PU traffic application. We define a threshold for sensing-transmission trade-off that is optimal under the technical constraints, and exploit short time and long time opportunities left by PUs. In the simulation results, we demonstrate the effectiveness of the proposed scheme by comparing with the threshold-based sensing transmission structure for throughput gains and show the sensing-transmission threshold under various traffic applications (patterns). Moreover, we illustrate the effects of different sensing times on throughput for different traffic applications using some real traces.