Improving Sensing and Throughput of the Cognitive Radio Network

In a cognitive radio network, when primary user's (PU's) spectrum is periodically sensed, secondary user (SU) needs to wait during sensing interval, which causes interruption in SUs' transmission and significant reduction in achievable throughput. If continuous spectrum sensing is implemented, then a portion of a band is always unavailable for SUs' transmission, which also limits the throughput. In this paper, we propose two methods. With the first one, we achieve an increase in throughput, while SUs' transmission is uninterrupted, and in the second method, we let our algorithm dynamically choose time-band portion of a frame for spectrum sensing, causing minimal interference to PU than the one in first method and offering uninterrupted service to cognitive radio (CR) users. Simulation results show that achievable throughput is improved significantly in both methods. It also showed that the methods offer normalized throughput between 4.7 and 5.3 bits/s/Hz over a wider range of sensing band from 0.75 to 6 MHz, when PU's active phase probability is 0.2, whereas delay-oriented continuous spectrum sensing (DOCSS) scheme achieves this range within a narrower sensing band--especially at the optimal value. Results also showed that for an arbitrary sensing band, normalized throughput achieved by our methods is larger than the one achieved by DOCSS scheme.