Modeling Periodic Sensing Errors for Opportunistic Spectrum Access

Abstract

In opportunistic spectrum access, a secondary radio (SR) periodically senses for a vacant channel to transmit. Common performance measures of the adopted sensing algorithm include missed detection and false alarm probabilities. However, minimal effort is made to incorporate the primary radio (PR) spectrum activities in the design. This paper models the errors arising from the mismatch between the SR channel sensing period and PR ON/OFF spectrum activities. Since the status of the channel is known, only at the sensing epoch, the PR may become active again in between the SR sensing periods, even if the idle channel is correctly sensed. The SR transmission thus collides with the PR transmission, and we refer to such a scenario as Type II missed detection error (MDE). Four performance metrics which serves as an indication of the SR sensing performance are identified, namely, the probability of Type II MDE, probability of missed SR transmission opportunity, probability of successful transmission and probability of SR blocked access. Based on results from renewal theory, we first model these metrics and then study their impact on the SR for Markovian and non-Markovian distributed PR spectrum activities. We then study the SR performance tradeoffs for different sensing periods. Simulated results verify the correctness of the analytically derived expressions. These results are important for designing an efficient SR system in the future.