Cognitive radio (CR) is regarded as a viable solution to enabling flexible use of the frequency spectrum in future generations of wireless systems by allowing unlicensed secondary users (SU) to access licensed spectrum under the specific condition that no harmful interference be caused to the licensed primary users (PU) of the spectrum. In practical scenarios, the knowledge of PU activity is unknown to CRs and radio environments are mostly imperfect due to various issues such as noise uncertainty and multipath fadings. Therefore, important functionalities of CR systems are to efficiently detect availability of radio spectrum as well as to avoid generating interference to PUs, by missing detection of active PU signals.;Typically, CR systems are expected to be equipped with smart capabilities which include sensing, adapting, learning, and awareness concerned with spectrum opportunity access, radio environments, and wireless communications operations, such that SUs equipped with CRs can efficiently utilize spectrum opportunities with high quality of services. Most existing researches working on CR focus on improving spectrum sensing through performance measures such as the probabilities of PU detection and false alarm but none of them explicitly studies the improvement in the actual spectrum utilization. Motivated by this perspective, the main objective of the dissertation is to explore new techniques on the physical layer of dynamic CR systems, that can enhance actual utilization of spectrum opportunities and awareness on the performance of CR systems.;Specifically, this dissertation investigates utilization of spectrum opportunities in dynamic scenarios, where a licensed radio spectrum is available for limited time and also analyzes how it is affected by various parameters. The dissertation proposes three new methods for adaptive spectrum sensing which improve dynamic utilization of idle radio spectrum as well as the detection of active PUs in comparison to the conventional method with fixed spectrum sensing size. Moreover, this dissertation presents a new approach for optimizing cooperative spectrum sensing performance and also proposes the use of hidden Markov models (HMMs) to enabling performance awareness for local and cooperative spectrum sensing schemes, leading to improved spectrum utilization. All the contributions are illustrated with numerical results obtained from extensive simulations which confirm their effectiveness for practical applications.
展开▼
