Queue performance of cognitive radio networks with general primary user activity model

The quality of service (QoS) performance analysis is of great guiding significance for a QoS guarantee of secondary users (SUs). However, the QoS performance of SUs has not been well studied, especially how it is impacted by the traffic property of primary users (PUs). In this study, the authors propose a method to analyse the queue performance of the SU when the active and inactive durations of PUs follow general distributions. To characterise the non-memoryless property of the channel when the distributions of PUs active/inactive durations are general distributions, they propose a two-dimensional Markov chain to model the states of the channel. By using this Markov chain, they derive the effective capacity (EC) function of the cognitive radio network. On the basis of the EC function and the effective bandwidth function, the queue performance of the SU, that is, the stationary tail distribution of the queue length, is estimated. The author´s result can not only be used to calculate other QoS metrics, such as the buffer overflow probability and throughput, but also provide some guidelines for QoS guarantee of the SU. Finally, they verify their work by comparing the analytical results with simulation results.