Spectrum And Energy Efficient Resource Allocation Techniques In Cognitive Radio Networks

Cognitive radio networks (CRNs) allow the unlicensed users to share the spectrum with the licensed users dynamically, and thus can alleviate the contradiction between the scarcity of available spectrum resources and the extremely underutilization of licensed spectrum. Therefore, CRNs are gaining more and more attentions from both academia and industry, which has been envisaged as one of the key network forms for future wireless communication networks. Aiming at achieving high system spectrum efficiency (SE) and energy efficiency (EE), how to exploit the resources such as spectrum and transmit power efficiently and dynamically while guaranteeing the quality of service (QoS) of licensed users and promoting the system performance for unlicensed users has become one crucial research issue in CRN due to the scarcity of spectrum and the limited transmit power for unlicensed users.Therefore, with the aim of maximizing SE and EE in overlay and underlay CRNs respectively, in this dissertation we combine the cross-layer design, optimization theory and robust design to study the efficient resource allocation schemes for CRNs. Here, overlay CRNs allow secondary users (SUs) to firstly perform spectrum sensing and then transmit data only when the target channel is sensed as vacant. While underlay CRNs permit SUs to send data simultaneously with primary users (PUs) provided that the interference incurred by SUs is tolerable for PUs’ transmission. Then, the main achievements and results of this dissertation are summarized as follows:1. We have proposed a novel channel aware multi-channel cognitive MAC protocol (CAM-MAC) to improve the SE in a distributed overlay CRN. For multi-channel CRNs, the inconsistent transmission conditions of common control channel (CCC) and data channels and the underutilization of different data channels’heterogeneous capacity degrade spectrum efficiency severely. Focusing on this problem, Chapter 2 has proposed a novel CAM-MAC protocol, which consists of a two-level four-way handshaking mechanism. On CCC, the first level two-way handshaking, which can both reserve data channels and convey the idle data channel formation, has been optimized to reduce the mean successful reservation time for SUs. On data channels, the other level two-way handshaking has been implemented via the cross-layer design approach to adopt the adaptive transmission based on the instantaneous signal-to-noise ratio (SNR) of each Nakagami fading data channel, which can fully utilize the heterogeneous capacity on different data channels. Finally, simulation results validate our analytical results and manifest that CAM-MAC outperforms the existing protocol with significant improvement in the saturated throughput and has better mean packet delay performance.2. We have devised an optimal power allocation strategy to maximize the mean EE for SUs in an underlay CRN. Focusing on the problem that the static optimization based on instantaneous channel state information (CSI) cannot promote SU’s EE and guarantee PU’s QoS over all the fading states in the fast fading underlay CRNs, Chapter 3 has investigated the power allocation scheme for SUs to maximize the mean EE while guaranteeing the QoS of PUs. Firstly, we have adopted the PU outage probability constraint as the QoS metric for PUs while taking both the peak and mean transmit power constraints for SUs into account, and then have formulated this problem as a fractional programming. However, this problem is nontrivial since the mean EE is non-convex and PU outage probability constraint belongs to chance constraints. With the aid of the fractional programming and Lagrangian duality theory, we have proposed an efficiently iterative power allocation (IPA) algorithm to derive the optimal power allocation strategy for SUs to maximize the mean EE while guaranteeing the QoS of PUs. Then, we have analyzed the computational complexity of IPA algorithm. Furthermore, we have found that SUs’SE maximization in this situation is a special case of SUs’mean EE maximization. Finally, simulation results assess the performance of our proposed power allocation scheme. Additionally, the tradeoff between SUs’ mean EE and PU outage probability threshold only occurs in some certain range.3. We have designed a robust power control scheme to maximize SUs’ EE in an underlay CRN with imperfect CSI occurring in all channels of interest. Imperfect CSI in underlay CRNs can reduce SUs’EE and impair the QoS of PUs’transmission severely. Focusing on how SUs can maximize the EE performance while guaranteeing the QoS of PUs’ transmission in presence of CSI errors in an underlay CRN, Chapter 4 has taken all channels’ uncertainties into account and formulated this problem as the max-min problem with infinite constraint from the perspective of robust optimization. However, this problem is nontrivial, because its outer-maximization problem is non-convex and its inner-minimization problem is a concave minimization problem known as NP-hard in general. Thus, according to the fractional programming and global optimization techniques, we have proposed an alternating iterative power allocation algorithm (AIA) to handle this problem. Then, simulation results validate that compared with the Non-Robust scheme, our proposed Robust method can improve the worst-case EE of SUs distinctly and strictly guarantee the QoS of PUs under all channel parameters’ uncertainties. More importantly, we have demonstrated that SUs’EE under Non-Robust scheme is robust to small channel errors with the large PUs’ interference threshold. Additionally, the EE under both schemes keeps nearly unchanged when the PU interference threshold is greater than a critical value.