Study On Capacity And Quality-of-service In Cognitive Radio Networks

Due to the rapid development of wireless technologies, the scarcity of spectrumresource becomes a serious problem for next-generation wireless networks. However,recent researches have shown that the allocated spectrum bands are severelyunderutilized. As a potential way to solve this problem, cognitive radio, has drawn a lotof attentions in the past decade.In this dissertation, from the perspectives of information theory, queuing theory andpercolation theory, the performance of cognitive radio networks are mainly investigatedin terms of two aspects: the capacity and the Quality-of-Service (QoS). For capacityanalysis, the capacity of small cognitive networks by taking consideration of theinfluence from primary users’ activity, the effective capacity with delay QoS constraintsand the scaling laws in interweaved ad hoc cognitive networks are investigated,respectively. For QoS analysis, some important QoS parameters are derived incentralized cognitive networks. The main contents of this dissertation are summarizedas follows:In Chapter1, the research background of our work is introduced, related researcheson cognitive radio networks are summarized, and the main contributions and contents ofthis dissertation are listed.In Chapter2, the capacity of cognitive radio networks with a primary user and asecondary user is investigated. Existing works are all based on the assumptions thatprimary activity does not change between two secondary sensing periods. However, inpractical scenarios, primary user may begin or finish its transmission during this period.In Chapter2, the influence of this event is considered. By deriving the capacity ofsecondary user, the relationship among primary activity, secondary sensing duration andcognitive capacity is discovered. The results can be used for frame designing ofsecondary user. Furthermore, the secondary packets transmission delays in the networkwith both Poisson arrival process and bursty traffic are also studied based on queuingtheory. In next generation wireless networks, all users will communicate with specific QoS requirements. Therefore, the capacity of secondary user with a QoS constraint incognitive networks is investigated, which is referred as an effective capacity problem.Unlike existing works in which both secondary transmitter and receiver have perfectchannel side information, only sensing result is known to secondary transmitter in ourmodel. The effective capacity is analyzed with respects to delay constraints and sensingduration. The results can be used to analyze cognitive network performance with QoSconstraints.The capacity of cognitive ad hoc networks is under hot discussion recently. However,the existing researches are all based on underlay cognitive networks, and the capacitiesof interweave cognitive networks are only discussed in small scale networks, i.e., anetwork with only one primary user and one secondary user. In Chapter3, we extendedthe study of information theoretic limits of interweaved cognitive networks from smallscale networks to large scale networks. By constructing two auxiliary networks, theupper bounds of the throughput of per primary user and per secondary user are obtained,respectively. By using the percolation theory, a cooperative relay scheme for this type ofcognitive network is proposed. It has shown that the achievable throughput in ourproposed scheme can meet the upper bounds of the network. It has shown that primaryusers can obtain higher transmission rates in interweaved cognitive networks thanunderlay cognitive networks, which may become a strong motivation for primary usersto share their spectrum resource.In Chapter4, an analytical model for evaluating the performance of centralizedcognitive networks is proposed, where the spectrum sensing is imperfect. Comparedwith existing researches, our model is more suitable for a multi-channel network.Moreover, the influence of sensing mistakes in spectrum handoff process is alsoconsidered. Thus, our model is close to reality more. The expressions for someimportant system QoS parameters are derived, such as, the primary (secondary)blocking ratio, the primary (secondary) forced termination ratio, and the spectrumutilization ratio. The results show that spectrum utilization improves significantly byintroducing cognitive radio technology, as well as by controlling the secondary traffic.Moreover, a certain QoS requirement of primary network may be guaranteed. Finally, the dissertation is concluded in Chapter5. Some suggestions for futureworks are also given in this chapter.