A new framework for dynamic spectrum access in cognitive radio networks

The radio frequency (RF) spectrum is a limited and precious resource. The continuing deployment of a diverse range of new wireless services into an already crowded RF spectrum requires a new way to accommodate these services. This increasing demand for the spectrum is in contrast with the inefficient use of the RF spectrum in some licensed bands, which high ights an opportunity for accessing the RF spectrum with cognitive radio (CR) networks. In CR networks, a secondary user group (the unlicensed users) is allowed to opportunistically access temporarily unused licensed bands of the primary users (licensed users), i.e., the spectrum holes. Accessing the spectrum band in such a manner poses many research challenges including finding the band vacancies, protecting the primary system from interference, allocating vacant bands among the secondary users, and sharing the spectrum bands among multiple CR networks. To address these issues, this dissertation proposes a new dynamic spectrum access framework for CR networks, including methods for the spectrum sensing, agile spectrum evacuation, spectrum allocation, and spectrum sharing. For spectrum sensing, a weighted combining scheme is introduced that intelligently assigns weights to the energy measurements of a number of CRs and then combines these values to decide if the primary user is currently using a band. This scheme exhibits a better detection accuracy and spectrum utilization. The agile spectrum evacuation scheme allows uninterrupted use of a licensed band by a CR until the primary transmitter returns when the CR quickly evacuates. This is a major innovation because until now a CR must interchange sensing and transmitting in a licensed band, creating a trade-off between spectrum utilization and unavoidable interference caused to the primary user. To allow the fair and fast allocation of available bands among CRs a collaborative framework for allocating multiple bands simultaneously among multiple CRs is proposed. This framework exhibits a significant superiority over conventional approaches in terms of an improved throughput and spectrum utilization, and reduced interference loss and collisions. Finally, a scheme to facilitate spectrum sharing among multiple CR networks is created by utilizing various game theoretic approaches for different configurations of band access. For each game, the Nash Equilibrium is defined and attained in most cases. Collaboration among the CR networks is also investigated through repeated games and it is shown that a cooperative method results in a much better sharing of the RF spectrum. Comprehensive performance analyses including mathematical formulations and experimental evaluations are provided, with the proposed dynamic spectrum access framework exhibiting a superior performance over existing techniques.