| Cognitive radio networks (CRNs) are a promising solution to the channel (spectrum) congestion problem. This dissertation presents work on the two main issues in CRNs: spectrum management and cross-layer protocol design.;The objective of spectrum management is to enable the efficient usage of spectrum resources in CRNs, which protects primary users' activities and ensures the effective spectrum sharing among nodes. We consider to improve the spectrum sensing efficiency and accuracy, so that the spectrum sensing cost is reduced. We consider the pre-phase of spectrum sensing and provide structures for sensing assistance. Besides the spectrum sensing phase, the sharing of spectrum, or the channel allocation, among nodes is also the main component in the spectrum management. We provide our approach to achieve a reliable and effective channel assignment.;The channel availabilities for different nodes in CRNs are dynamic and inconsistent. This poses challenges on the MAC layer protocols for CRNs. Moreover, due to the lack of knowledge on primary users, they can suddenly become available during the secondary users' data transmission. Therefore, for a end-to-end data transmission in CRNs, the routing algorithm is different from the existing routing algorithms in traditional networks. We consider the cross-layer protocol design, and propose the solutions for efficient data transmission. We propose the novel routing protocol design considering the boundaries of PUs. Also, an effective structure for reliable end-to-end data transmission is presented, which makes use of the area routing protocol. We build a USRP/Gnuradio testbed for the performance evaluation of our protocols. |
