A New Method for Channel Availability Analysis and the Associated Policy Design for Selection of Channel Sensing Order in CRNs

We consider the problem of the selection of channel sensing order in cognitive radio networks (CRNs) where the primary users (PUs) use multiple channels for data transmission. We assume that the occupancy state of each channel is correlated with that of the other channels, and the spectrum occupancy state for all the channels follows the discrete-time Markov model. In this paper, we propose to characterize the channel availability of the primary network using the idle probability of consecutive time slots (IPCOTS). Under the discrete-time Markov model, we derive the IPCOTS of each particular channel, under the condition that the states of all the channels in the primary network are identically distributed. Moreover, we propose a suboptimal state-dependent policy, called IPCOTS policy, for the selection of channel sensing order. Finally, we extend the policy to three typical cases. Simulation results show that, in an M/M/m/m-modeled primary network, more secondary user (SU) rewards can be obtained using the IPCOTS policy in comparison with the existing approaches.