Spectrum sensing aided long-term spectrum management in cognitive radio networks

Wireless microphones operating in the TV white spaces often appear at specific venues such as schools or churches, and at specific times. Hence, their location and appearance pattern can be predicted from spectrum sensing statistics. In this paper we propose and evaluate three spectrum selection functions that utilize sensing results to provide long-term spectrum usage statistics as basis for channel selection to enhance performance by reducing interference and increasing throughput. To evaluate performance of the spectrum selection functions, these are implemented in a detailed system level simulator for the IEEE 802.22 standard. We find that the spectrum selection function that uses statistics about channel idle and busy periods performs best when primary user activity is high, and that the spectrum selection function that uses predictions about location and distance to primary users performs best when IEEE 802.22 radio users are mobile and the primary user activity is low.