Cognitive radio (CR) has been considered as an innovative and powerful technology to solve the spectrum under-utilization problem in conventional wireless networks. Instead of fixed spectrum allocation, CR allows cognitive users to dynamically and intelligently share the radio spectrum with licensed users. Unfortunately, cognitive radio networks (CRNs) impose unique challenges due to the co-existence of various primary networks, of which the Quality of Service (QoS) should be protected. Most of the existing research on CR has mainly focused on detecting and utilizing unused portions of the spectrum (i.e., spectrum holes) in the frequency-time domains. However, spatial dimension spectrum reuse has not been carefully investigated. By exploiting location information, cognitive use...[ Read more ]

Cognitive radio (CR) has been considered as an innovative and powerful technology to solve the spectrum under-utilization problem in conventional wireless networks. Instead of fixed spectrum allocation, CR allows cognitive users to dynamically and intelligently share the radio spectrum with licensed users. Unfortunately, cognitive radio networks (CRNs) impose unique challenges due to the co-existence of various primary networks, of which the Quality of Service (QoS) should be protected. Most of the existing research on CR has mainly focused on detecting and utilizing unused portions of the spectrum (i.e., spectrum holes) in the frequency-time domains. However, spatial dimension spectrum reuse has not been carefully investigated. By exploiting location information, cognitive users have the potential to estimate the spectrum holes more precisely and optimize their power control, spectrum access and routing strategies so as to improve spectrum efficiency. Due to the high fluctuation and heterogeneity of the available spectrum occupied by licensed users, there are great challenges to develop location-aware dynamic spectrum sharing techniques, not only in the physical layer but also across the whole layers of the remaining protocol stack.

In this thesis, we investigate various location-aware design approaches, including location-aware spectrum management, location-aware network optimization and location-aware routing protocols, to exploit multi-user, multi-frequency, and multi-hop selection diversity. We start with investigating the spectrum access for underlay CRNs. The transmit power is regulated to identify the spatial spectrum holes. To take advantage of the distributed spectrum holes, a short-path routing strategy is developed for multihop CRNs. Such a strategy navigates the transmitting packet from the source to the destination by following a designed guide strip, while avoiding harmful interference to the primary users. Next, a joint power control, spectrum allocation and relay selection scheme is considered for multiuser CRNs, which further improves the spectrum and power efficiency by coordinating the spectrum holes on multiple frequency bands. To fairly allocate the spectrum, we then design a novel semi-matching spectrum sharing framework with the objective of maximizing the minimum throughput among the cognitive users. With graph theoretic tools, efficient and effective algorithms are proposed based on knowledge of the location information. By numerical results, these designed algorithms are demonstrated to enhance the cognitive users' QoS, as well as to improve the spectrum efficiency and fairness.