Analytical Framework for Throughput Evaluation of Practical Opportunistic Spectrum Access Networks: A Cross-layer Approach

Opportunistic Spectrum Access (OSA) network operates on a temporary idle spectrum that is already assigned to primary users (PU), allowing secondary users (SU) to overcome spectrum scarcity problem. In general, OSA system detects available spectrum by reliable spectrum sensing schemes and utilize the detected spectrum for data transmission.;While various effective methods for spectrum sensing or resource utilization are studied in literature, there is still a lack of research on a comprehensive framework for throughput evaluation in order to compare the differences in performance improvements of various communication methods. In this thesis, we develop an analytical framework to evaluate system throughput of practical OSA network. First, we analyze spectrum sensing and resource utilization strategy into layers, such as radio layer, physical layer (PHY) and medium access control layer (MAC). Then, capturing interaction between layers, we derive throughput as a product of time efficiency and channel utilization. The developed analytical framework is able to capture the interplay between spectrum sensing and resource utilization. For example, it evaluates the relation between the idle spectrum detected by spectrum sensing and the performance of resource utilization scheme. In order to analyze practical networks, we apply this framework to ad hoc and centralized cases.;First, we present the analysis of the OSA ad hoc network. In order to achieve high throughput for an ad hoc network where nodes are distributed over multiple channels, we consider various combinations of collaborative spectrum sensing and multichannel data Medium Access Control (MAC) protocol abstractions. For collaborative sensing, we parameterize each layer, classify existing solutions, and propose a new protocol called Truncated Time Division Multiple Access (TTDMA) that supports efficient distribution of sensing results in "k out of N" fusion rule. In case of multichannel MAC protocols we evaluate two main approaches of control channel design with (i) dedicated and (ii) hopping channel. We propose to augment these protocols with options of handling secondary user (SU) connections preempted by primary user (PU) by (i) connection buffering until PU departure and (ii) connection switching to a vacant PU channel. By comparing and optimizing different design combinations we show that (i) it is generally better to buffer preempted SU connections than to switch them to PU vacant channels and (ii) TTDMA is a promising design option for collaborative spectrum sensing process.;Second, we apply the developed framework to the centralized network. Because centralized network needs to be managed by a base station, complex and efficient methods, e.g. OFDMA and two-stage sensing, are possible. Thus, we derive an analytical model that enables a comparison of multiple design options of Opportunistic Spectrum Orthogonal Frequency Division Multiple Access (OS-OFDMA). The model considers continuous and non-continuous subchannel allocation algorithms, as well as different ways to bond separate non-continuous frequency bands. Different user priorities and channel dwell times, for the Secondary Users and the Primary Users of the radio spectrum, are considered. Further, the model allows the inclusion of different qualities of services of Secondary Users, such as constant bit rate (CBR) and variable bit rate (VBR) services. Then, the model considers an efficient two-stage spectrum sensing algorithm that is designed to reduce sensing time overhead. From the analysis we conclude that for the case when large number of NPUs operates with high activity, OS-OFDMA with subchannel notching and channel bonding could provide almost an order of magnitude higher throughput than the design without those options enabled.;In summary, this is the first general framework which can provide comprehensive throughput evaluation for specific and practical design options of sensing and multichannel data MAC of OSA network. Using this framework, we were able to make distinctive contributions exemplified as follows: (i) performance comparison of connection buffering and connection switching; (ii) derivation of parameters of collaborative sensing that maximize throughput; (iii) development of a new efficient protocol for collaborative sensing, TTDMA; (iv) evaluation of throughput for subchannel notching and channel bonding schemes for OS-OFDMA; and (v) analysis of two-stage sensing algorithm. Conclusively, for any new technology for OSA network, we can use the developed framework to evaluate the throughput and present guidelines for the system design.