Spectrum Sensing And Spectrum Access In Cognitive Radio Networks

As a kind of new intelligent wireless technology, cognitive radio provides a feasible solution for the scarcity problem of spectrum resources. Different from traditional wireless networks, cognitive radio networks allow dynamic spectrum sharing between cognitive users and primary users, which realizes high efficient and flexible spectrum utilization. Spectrum sensing and spectrum access are studied in this thesis. Spectrum sensing makes the cognitive users be able to detect the licensed band owned by primary users and obtain the information of current spectrum utilization. Then, all available spectrum holes are searched and found. Spectrum access helps the cognitive users make analysis and selections with the available bands obtained by spectrum sensing, so as to achieve high-efficiency, reliable spectrum access, and get the spectrum using permission.First, the thesis gives a brief introduction of the cognitive radio. Then, the related key technologies and main performance evaluation indexes are summarized.Second, taking into account that the states of the primary user have time correlations between two adjacent sensing periods, an improved sensing algorithm with decision trust strategy is proposed. In this algorithm, a sliding window is set to estimate the primary user’s state transferring probability. Then, taking minimizing the error probability as the optimization criterion, three different situations in decision trust strategy are deduced based on first order Markov model. The simulation results show that the proposed spectrum sensing algorithm has lower error detection probability and the sensing performance has a certain degree of improvement.Third, considering to the fact that the diversity of services has influence on spectrum access, all secondary users are divided into two types based on bearer services:delay sensitive secondary users(S-SUs) and delay tolerance secondary users (T-SUs). Then, a six states continuous time Markov model is built for cognitive systems. A service-based spectrum access strategy is proposed to achieve different ways to access the band for different QoS demands. For S-SU, they do spectrum access without buffer and with high access probability to ensure high access priority; For T-SUs, a buffered spectrum access method is provided for them to improve their one-time successful access probability. The simulation results show that the service based spectrum access strategy can achieve better performance (such as transmission throughput and channel utilization efficiency). In addition, there exists optimal access probabilities for S-SU and T-SU which makes the cognitive network well balanced between the transmission performance and the fairness among users.Finally, the conclusions of this thesis are present. Future study directions are also discussed in the last part of this thesis.