Research On Spectrum Sensing And Resource Allocation Optimization In Cognitive Radio

With the development of wireless communication technology, the existing spectrum resource has become more and more crowded. In addition,fixed spectrum allocation policy also led to low spectrum utilization. As a key technology to solve this problem, cognitive radio has been received more and more attention. This dissertation focuses on spectrum sensing and optimal resource allocation in cognitive radio, respectively. The asynchronous cooperative spectrum sensing, the physical security threat, and joint optimization with sensing time and spectrum allocation are studied with theory analysis and simulation experiment.In synchronous cooperative spectrum sensing, a synchronous clock is required in the sensing and reporting processes, but it is difficult to synchronize straightly and synchronous clock would cause large overhead. Therefore, it is necessary to study the asynchronous cooperative spectrum sensing. Aimmed at sloving the problems of bad timeliness and large computation in current asynchronous cooperative spectrum sensing schemes, a triggered asynchronous cooperative spectrum sensing scheme based on Dempster-Shafer(D-S)theory is proposed. Sensing asynchronously, each cognitive user calculates the confidence measure functions with the double threshold spectrum sensing method, and then transports their sensing results to the fusion center. Receiving useful sensing information, with the altered-result triggered nature, fusion center uses D-S theory to fuse the sensing results. The analysis and simulation results show that this method can improve the asynchronously spectrum sensing efficiency and reduce the overhead of the cognitive system.Cognitive users are assumed to be honest and cognitive users’ local sensing results are treated equally in most cooperative spectrum sensing schemes. In practical cognitive radio networks, not all the cognitive users are honest and malicious user may appears in the cognitive radio network which can counterfeit the local sensing information and send false sensing results to the central fusion center, leading the fusion center make the wrong decision. In order to solve this problem, an improved malicious users detected scheme based on belief degree is proposed, in which stackelberg game is adopted to improve the sensing performance. According to the reliability of cognitive users, the proposed scheme can make a distinction between leaders and followers. Considered as leaders, users with acceptable reliability will share their sensing observations with the followers. According to the reliability difference between the malicious users and other users, the proposed scheme can identify many types of malicious users easily.When the secondary user occurs missed detection,the secondary user at farther distance from the primary user can potentially incur lower interference on the primary user compared to one at closer distance. Therefore, only considering probability of missed detection may overly protect the primary user. In order to solve this problem, this dissertation first introduces an interference-aware spectrum sensing model defining the probability of interference as a new metric for spectrum sensing, then this dissertation studies the optimal resource allocation strategy that maximizes throughput of cognitive system under transmit power constraint of secondary user and the interference power constraint of primary user as well as the new metric-interference probability jointing missed detection probability constraint, and establishes the mathematical model of multi-constraint conditions, then solves the optimal sensing time and resource allocation by developing a subgradient method in convex optimization theory. Finally, simulation results show the feasibility and superiority of this scheme.There are 49 figures, 8 tables and 114 references in this dissertation...