Research On Effective Capacity And Power Control Of Cognitive Wireless Networks

The cognitive radio (CR) networks allow cognitive users or the so-called secondary users(SUs) to communicate over certain bandwidth originally allocated to a primary user.Generally speaking, there are three basic operation models for CRs: underlay, overlay, andinterweave. In underlay model, SU is allowed to transmit as long as the interference from SUdoes not degrade the quality of service (QoS) of primary users(PUs) to an unacceptable level.SUs should control their transmit power properly in order to achieve a reasonably hightransmission rate without causing too much interference to PUs. Accordingly, dynamicresource allocation (DRA) for the CR becomes crucial, whereby the transmit power levels,bit-rates, bandwidths, and/or antenna beams of the CR are dynamically changed based uponthe channel state information (CSI) across the primary and secondary networks. As a result,SUs achieve the best performance, while maintaining a required quality of service for eachcoexisting active primary link.The increasing demand for wireless network services such as video/audio over IPmotivates an unprecedented revolution in wireless broadband networks. Due to the spectrumsharing and interference power constraints, this can be more challenging to satisfy delayconstraint in CR networks. Therefore, it is very important to research QoS-driven powercontrol scheme in CR networks.In this paper, the QoS-driven power control schemes are proposed based on effectivecapacity model in spectrum-sharing CR network. The relationships between the optimizedparameters and delay constraints are investigated for different fading environments. Wepropose a scheme of increasing the effective capacity by diversity-based CR systems, and ascheme of decreasing interference power to the PU by OFDM-based CR systems for the delaysensitive traffics. The main contributions are as follows:(1) Firstly, a QoS-driven power control scheme is proposed for single antenna CRsystems. The proposed scheme aims at maximizing the system throughput subject to a givenstatistical delay-QoS constraint and interference temperature threshold. The expressions ofpower control strategy are derived for the most loose delay constraint and the most stringent delay constraint. We derive the closed-form expressions for the effective capacity underdifferent fading channels. We investigate the impact of asymmetric fading channels on theeffective capacity with delay QoS guarantees. More specifically, we allow thesecondary-to-primary user and secondary-to-secondary user channels undergoing differentfading types and arbitrary link powers. Our main results indicate that the effective capacity issensitive to the fading types and link powers. For the looser delay constraints, the interferencelink fading parameters play a vital role. While for the more stringent delay constraints, thecognitive link fading parameters play a decisive role. And on this basis, we extend our studyfor one primary user to multiple primary users. The impact of multiple primary users on thecapacity gains under delay constraints has also been explored. We observe that the effectivecapacity degrades with increasing number of the PUs for the looser delay constraints, whilethis influence disappears gradually for the more stringent delay constraints.Secondly, we explore the relationships between he effective capacity and delay QoSconstraint for different diversity branches for diversity-based systems. Our results indicatethat increasing the antenna diversity branches can improve the cognitive capacity of channelperformance. Especially for the stringent delay constraints, the effective capacity gains aresignificantly improved when multiple antenna branches are deployed.(2) Firstly, a QoS-driven power control scheme is proposed for single-carrier CR systems.The proposed scheme aims at minimizing interference power to PU from SU for a certainarrival traffic flow, such that a required delay-QoS constraints can be satisfied. Theclosed-form expressions of average interference power are derived for Nakagami fadingchannels. The relationships between interference power and delay QoS constraints areinvestigated. Our main results indicate that there have to trade off the interference power forQoS provisioning for single carrier CR systems. For the stringent delay constraints, theinterference power is greater than for the loose delay constraints. When the SU-to-SU channelis in deep fading, the interference power is increasing sharply with the increasing of QoSexponent. When the quality of SU-to-SU channel is better, the interference power isincreasing slowly with the increasing of QoS exponent.Secondly, a QoS-driven power allocation scheme is proposed for multi-carrier CR systems.The proposed scheme aims at minimizing total interference powers for a certain arrival traffic flow, such that a required delay-QoS constraints can be satisfied. Over the frequency domain,the proposed allocation is always water filling, regardless of the QoS constraint. Over the timedomain, the proposed allocation depends on QoS constraint. we explore the relationshipsbetween the total interference powers and delay QoS constraints for different number ofsubchannels. Our results indicate that multi-subchannels are deployed in order to reduce theinterference powers for the stringent delay QoS constraints（3）We study the effective capacity gains of CR wireless systems with maximal ratiocombining (MRC) diversity at the secondary receiver. The SU does not have perfect channelinformation of SU-to-PU links, where estimation error based on minimum mean squareerror(MMSE) is considered. Considering average interference power constraint, we derivethe effective capacity expression along with its optimum power control policy. Our resultsshow that the effective capacity decreases while estimation error increases. MRC not onlycould improve capacity gains, but also could compensate the estimation errors.The effective capacity is investigated considering interference outage constraint withimperfect channel information. We derive lower bounds of effective capacity aboutinterference outage power control policy. The relationships among effective capacity,interference outage and estimation error are explored. Our results show that even a smallamount of interference outage permitted by the PU, results in a significant gain in effectivecapacity compared to a system with strict peak interference power constraint. Especially forlooser delay constraints. In addition, compared with average interference power constraints,the capacity loss as an effect of imposing a peak interference power constraints is notsignificant as long as some level of interference outage. This is completely different comparedto perfect CSI.