Perception Network Theory And Algorithm Of Dynamic Spectrum Sharing

With the rapid development of wireless communication systems, the problem of spectrum resources scarcity is becoming more and more urgent. In order to figure out this problem, people proposed the cognitive radio technologies. This kind of technology enables wireless communication devices to independently learn about the outside environment and adaptively adjust their working status to achieve the best communication performance. The three basic functions of the cognitive radio technology are spectrum sensing, spectrum mobility management and spectrum sharing. Spectrum sharing allows multiple cognitive/primary users to efficiently share the spectrum resources. It is the key technology to improve the spectrum efficiency of cognitive radios. This thesis mainly focuses on the spectrum sharing technology in cognitive radio networks.This thesis first introduces the reader to the definition, background and basic idea of cognitive radio. Second, the paper overviews the three basic functions of cognitive radio. The architecture and application of cognitive radio networks are also briefly introduced. Based on the above contents, the thesis extensively presents the main research content of dynamic spectrum sharing technology, including medium access control, distributed spectrum sharing and power control.The spectrum sharing system must limit the power of interference imposed on primary users by secondary users, meanwhile the secondary users try their best to enhance throughput performance of themselves. Hence, power control of secondary users is very important. Some existing works utilize the competitive market equilibrium model to describe this kind of problem and conduct some preliminary research work based on that model. This thesis first identifies the difference between the competitive market equilibrium problem and the linear complementary problem. Second, we derive the analytical solution to the simplified version of the competitive market equilibrium problem. Finally, inspired by the tatonnement process, we propose a hybrid algorithm. This algorithm can fully utilize the tatonnement process to converge to the competitive market equilibrium point. Meanwhile, when the tatonnement process diverges, the proposed algorithm automatically switches to an alternative power allocation scheme.Dynamic spectrum access is one of the key technologies to implement dynamic spectrum sharing. This thesis proposes a dynamic spectrum access algorithm in distributed cognitive radio networks which can be used in the case that no control channel is available. The proposed algorithm can be an effective complementation to the dynamic spectrum access algorithms based on control channels. The proposed algorithm divides the secondary users into active and passive users and establishes a one-on-one mapping between the members of the two groups. The proposed algorithm also adopts a deterministic channel selection mechanism to coordinate the channel allocation of secondary users without the need of control channels. Compared with similar algorithms, the proposed algorithm has some advantage on aggregate throughput performance and improves the fairness of channel allocation of secondary users as well. This conclusion is verified through computer simulation results. Moreover, existing algorithms often invoke the time-consuming pseudo random number generators to determine the nodes’status and working channels. The proposed algorithm does not use any pseudo random number generator and thus has the potential to reduce the processing delay.In general, the dynamic spectrum sharing is implemented in the medium access control (MAC) layer. This thesis proposes a MAC protocol in distributed cognitive radio networks. The proposed MAC protocol utilizes a channel selection algorithm which enables secondary users to detect the channel with the lowest busy probability and reduces the interference to the primary users. Second, the transmitting and receiving nodes shake hands to exchange channel availability information, reducing the spectrum sensing errors. At last, each cognitive node utilizes a dedicated control transceiver to exchange control messages over the control channel, solving the multi-channel hidden terminal problem. The contributions of this thesis are three-fold:the proposed cognitive radio MAC protocol provides Quality of Service (QoS) guarantee mechanism and can provide differentiated services to different types of users; a sensing channel selection mechanism is designed. This mechanism does not use a priori knowledge on the statistics of primary users and is easy to be implemented; we theoretically analyze the performance of the proposed MAC protocol in both the saturated and non-saturated network cases. Through sufficient computer simulations, the performance advantage of the proposed protocol is assessed and the accuracy of the theoretical analysis is also verified.