Wireless access through unlicensed spectrum is a major component for future broadband initiatives. Unfortunately, unlicensed frequency bands are congested, and existing technologies do not provide quality assurance for users with different access priorities and needs. The lack of quality of service guarantee is a key roadblock to providing multimedia content and facilitating broadband communications. There is a cogent need for technologies that provide distributed hierarchical access of unlicensed spectrum through dynamic spectrum sensing, learning, and cognitive sharing of spectrum opportunities.
This research develops mathematical theory and practical methodologies for the cognitive access of wireless spectrum shared by multiple classes of users. The technical approaches of this research are based on the structural optimality of certain carrier sensing policies. The optimality and simplicity of these policies result in new practical yet optimal solutions to cognitive spectrum sharing among users from different priority classes. The overall objective is twofold. First, it aims at gaining a foundational understanding of the limits of cognitive spectrum access in hierarchical networks. In particular, it characterizes the maximum throughput and effective bandwidth regions for users for different priority classes by exploiting structures of optimal policies for cognitive sensing and access. Second, this research develops engineering practical solutions for decentralized cognitive access of shared spectrum. In particular, new multiuser sensing and access protocols are developed and tested for practical applications. Effects of fading, interference and other implementation issues are also examined. This research includes a significant education component aimed at integrating frontier research with undergraduate and graduate curricula.
This research develops fundamental theory and practical techniques for the cognitive access of wireless spectrum shared by primary, secondary, and possibly multiple classes of users. The developed technology is an essential component in capturing spectrum white space, allowing users from a lower priority class to access spectrum when the users in the higher classes are inactive, thus enhancing spectrum utilization efficiency. This project achieves a dual objective. First, it obtains foundational understanding of the limits of cognitive spectrum access involving multiple classes of users. This includes the characterization of maximum throughput region of the secondary network. When the secondary users have additional queueing delay requirements as part of their Quality of Service (QoS) specifications, the maximum region of effective bandwidths will be characterized. The project also obtains fundamental tradeoffs among key design parameters and examines structures of optimal sensing and access policies. Second, this project derives engineering solutions for a decentralized cognitive access of shared spectrum. In particular, the project results in new multiuser sensing and access protocols for practical applications, providing better understanding of effects of fading, interference, and other implementation issues. The project has broader impacts in areas important to society. As an engineering research project, the results from this project have technological and economical impacts on the utilization of the spectrum resource. It supports a new access paradigm that separates users into different service classes with economical incentives to share surplus transmission opportunities. The developed technology also has potential impact on the new paradigm of cloud computing as it provides the critical link between local computation and resources in the cloud.
