Research On Cooperative Spectrum Sensing And Performance Optimization In Cognitive Radio Networks

With the increasing development of future mobile communications, to meet people’s desire for explosive growth of data traffic and continuous emergence of new services, the fifth-generation (5G) mobile communications system emerges at the right moment. Mobile Internet and Internet of Things (IoT), as the two main drivers in the future development of mobile communications, have attracted considerable research interest. As a key technology for future mobile Internet and Internet of Things, cognitive Radio (CR) can sense the spectrum environment and process detection, analysis, learning and planning to adaptively adjust its own parameters, for the purpose of realizing dynamic spectrum access and enhancing the spectral efficiency. Considering network diversity for future mobile communications, the design of effective resource allocation schemes for different frame structures in CR networks to further improve the spectral efficiency, has become a research focus in academia. Therefore, considering the imperfect channel state information (CSI) and relay nodes introduced in CR networks, it is of great significance, theoretically and practically, for us to design effective algorithms to improve the performance of existing cognitive radio networks in combination of new technologies to solve the optimization problem with the optimization conditions diversified and optimization variables multi-dimensional.Aiming to improve the sensing performance and capacity of existing CR networks based on different frame structures, this dissertation combines new technologies, such as relay and simultaneously wireless information and power transfer (SWIPT), to study the resource allocation algorithms. To begin with, the optimization design of resource allocation schemes in one-layer CR networks, including the distributed CR network and the centralized CR network, are studied. Then, the research turns to two-layer hierarchical CR networks, thejointoptimization of clustering algorithm and weighted data fusion to respectively maximize the system sensing performance and the network throughput are investigated. Finally,the resource allocation schemes are proposed for the energy-constrained cognitive relay network. The main work and contributions of this dissertation are summarized as:Firstly, this dissertation focuses on the problem of resource allocation for cognitive relay networks and proposes a novel N+1-phase transmission protocol and the optimization algorithm for relay power allocation.Considering a scenario where multiple SU source and destination pairs co-exist with a PU and a relay, we minimize the system outage probability subject to the power constraint as well as the interference to the PUs constraint. And then the model between the relay power allocation and the system outage probability is established. The closed-form solution is derived according to the optimization theory. Finally, the simulation results show that the proposed algorithms exhibits higher performance than the traditional algorithms.Second, the problem of frame structure optimization is addressed in distributed CR networks. Thedissertation first proposes a novel data transmission frame structure and establishes the trade-off model between the throughput and iteration numbers. Then, the optimal throughput is demonstrated to be achieved with respect to a limited iteration number. In additional, a new iterative algorithm is proposed according to the above derivation. Finally, simulation results show the proposed scheme offers superior performance in terms of the average system throughput and inferior complexity in term of iteration number.Third,considering both the perfect and imperfect channel state information, thejointoptimization of clustering algorithm and weighted data fusion rule to respectively maximize the system sensing performance and network throughput in the two-layer hierarchical CR networks are inveatigated. Specifically, three sets of variables, i.e., detection threshold, the clustering algorithm and different weights, are considered in our optimization problem. Obviously, it is a mixed integer programming problem under constraints. To simplify the solution, reasonable detection threshold and weight fusion rule are first optimized. Thus the problem is transformed as a constrained discrete optimization problem. Finally, numerical results show that the proposed scheme achieves a satisfying performance compared to traditional K-mean scheme.Fourth, the problems of resource allocation in cognitive radio networks with wireless energy harvesting are addressed. We first consider a three-phase energy harvesting and information transmission protocol. The objective is to maximize the ergodic capacity of the primary system and the secondary system. And the formulated optimization problem is a combinatorial problem involving both discrete and continuous variables. To simplify the solution, a reasonable relay selection is first proposed. And then the power allocation algorithm is studied. Finally, numerical results show that the proposed schemes achieve a satisfying performance.Finally, conclusions and future work are summarized in the end of this dissertation.