A dynamic forecast scheme of spectrum occupancy for opportunistic spectrum access in cognitive radio

In opportunistic spectrum access (OSA), not all the spectrum holes detected by Secondary System (SS) are available for SS users. Spectrum analysis is a required step before SS users access the detected spectrum holes in data slot. SS should make sure that the quality of service (QoS) of primary system (PS) users would still be acceptable after SS accessed the spectrum holes. This paper proposes a dynamic forecast scheme of spectrum occupancy based on the forecast of the QoS of PS users. The average queuing delay of PS users is used as an indicator of the impact of OSA on the QoS of PS. SS decides which spectrum holes SS user is going to access and how long the OSA will last based on the forecast of average queuing delay of PS users in data slot. The simulation results show the proposed scheme can reduce the risk of severely impacting PS performance in OSA, especially when the data slot is short.