Research On Cooperative Spectrum Sensing Technologies Based On Energy Detection

As one of the most promising technology that has the potential of dealing with the stringent requirement and the great scarcity of the radio spectrum, cognitive radio(CR) is attracting growing attentions these years. It allows secondary users(SUs) to dynamically access idle spectrum without generating harmful interference to primary user(PU) and thus can lead to higher utilization level for the available spectrum. Spectrum sensing is the fundamental of CR, and its nature is to observe whether PU employs the interested frequency or not. As common spectrum sensing methods without collaboration, such as energy detection, cyclostationary feature detection and matched filter detection, usually degrade with shadowing, multipath fading and noisy uncertainty, cooperative spectrum sensing(CSS) is proposed to combat these effects. Regardless of the improvement of the detection performance, CSS brings out new cooperation overheads including common control channel, reporting delay and energy consumption and so on. Therefore, this thesis focuses on CSS based on energy detection, and proposes the following solutions to achieve the target of improving cooperative sensing performance and decreasing sensing overhead:First, considering the CSS in flat Rayleigh fading channel, a novel user selection scheme is proposed based on finite state Markov channel(FSMC). By employing FSMC model to depict the variance of the fading channel, the switching rule between sensing and sleeping for SUs according to current status is designed, and two implementation methods are also provided depending on the availability of prior knowledge about channel parameter. As the research indicates, the proposed scheme can maintains SUs with low SNR mostly in sleeping while SUs with high SNR in sensing as much as possible, consequently realizes the purpose of user selection dynamically, and improving the cooperative sensing performance as well as reducing the sensing overhead.Then, concerning there are too many idle reporting slots in CSS, a novel frame structure is designed. In the designed frame structure, when an SU is reporting the sensing results, the following SUs continue local sensing until their turns to report. Considering the requirements of both constant false alarm rate(CFAR) and constant detection rate(CDR), the optimal schedules of SUs under AND and OR rules are investigated respectively, followed by the discussions of the sensing performance gains under the proposed frame structure. As the analyses show, since a half of idle reporting slots are utilized appropriately for sensing, the designed frame structure enhances the cooperative sensing performance considerately without any additional sensing delay, and thus behaves more efficiently.Next, aiming at the CSS conducted by SUs with energy constraint, a novel censoring scheme is designed, whose core idea is to let an SU report an indication signal only if the sensing result differs from its previous one. The reporting energy in the proposed scheme is investigated along with comparison, and then the spectrum sensing period is optimized subject to sufficient protection to the PU. As the analyses suggest, the proposed scheme reduces the reporting energy significantly without any loss in cooperative sensing performance. Moreover, in case of imperfect reporting channel, an improved scheme is proposed to guarantee the feasibility of our proposed censoring scheme.Finally, a sequential decision fusion(SDF) scheme based on conventional K/N rule is proposed, by sequentially receiving local decisions, the FC may achieve a global decision before receiving all individual decisions. We verify that the performance of CSS using the proposed scheme is exactly the same as the conventional CSS where FC adopts all decisions, then we examine the average number of decisions required at FC, followed by the closed-form under AND and OR rules. As the analyses show, the proposed scheme reduces the average number of decisions required considerately without sacrificing the cooperative sensing performance, and thus effectively saves reporting energy of the SUs. Additionally, a combined design of this scheme and the previous censoring scheme is considered as well, which further reduces the reporting energy.