Full-duplex cognitive radio with asynchronous energy-efficient sensing

Using a novel embedded Markov chain, we model and analyze a cognitive radio performing full-duplex spectrum sensing which is being carried out imperfectly-i.e., with errors-and asynchronously with primary traffic, from the perspective of energy efficiency. The effect of sensing frequency, which is varied by inserting sleeping periods between sensing processes is investigated, with focus on: 1) the energy efficiency of the device measured in terms of the number of successful transmissions under a limited battery budget; 2) the average throughput; and 3) the collision with the primary's traffic. We show analytically that, given false-alarm and mis-detection probabilities, the device's operation in lower-than-maximum sensing frequency may be more energy-efficient than that in maximum-frequency sensing case, while the radio is neither suffering throughput degradation nor disturbing the primary traffic seriously. We validate the deployment of such full-duplex cognitive radio (FDCR) along with the proposed sensing scheme for low-power short-range applications like wireless machine-to-machine communications and sensor networks, where the share of sensing power is comparable to that of transmission power. The merits of the proposed FDCR scheme are demonstrated through comparisons with a half-duplex cognitive radio scheme under different operating conditions and full-duplex self-interference cancellation factors.