Research On Dynamic Resource Allocation In Cognitive Relay Networks

With the the wide proliferation of intelligent terminals as well as the explosive growth of data traffic,there is an ever-growing demand for radio spectrum resource.Radio spectrum resource is a kind of non-regenerated resource.And traditional way to allocate spectrum resource is fixed and can cause underutilization of the spectrum resource.Cognitive Radio(CR)has been proposed as an intelligent wireless communication technology to solve the problem of spectrum scarcity and underutilization of the spectrum resource.In cognitive radio networks(CRNs),there are licensed users and unlicensed users,i.e.,primary users(PUs)and secondary users(SUs).SUs can access to the spectrum of PUs as long as guaranteeing the quality of service(QoS)of PUs,which can make full use of spectrum resource.Orthogonal frequency division multiplexing(OFDM)technology has been applied to CRNs due to the advantages of high spectrum efficiency,strong ability to combat multipath time delay,and capability to eliminate the inter-symbol interference.In OFDM-based CRNs,according to the channel feedback state information from surrounding equipments,SU transmitters(SU-TXs)dynamically allocate subcarriers and power to satisfy the interference thresholds of PU receivers(PU-RXs)and the demand of SUs.The SUs have to increase their transmit power to guarantee the reliablity of communication when there are the scenarios of long distance and severe channel attenuation between the pair of SUs,which can bring harmful interference to PU-RXs.Cooperative relay technique has been introduced into CRNs to extend coverage,increase spectrum efficiency and solve the problem of large transmit power caused by long distance.However,the complex topology of cognitive relay networks brings huge challenge to resource allocation.Considering the existing difficult problem and the trend of future development for cognitive relay network,this thesis focuses on the dynamic resource allocation problem in cognitive relay networks from several directions: channel state information(CSI)uncertainty,quality of experience(QoE),energy efficiency and physical layer security.The main contributions of this thesis are concluded as follows.First,most of the existing researches on resource allocation for cognitive relay networks are developed with perfect CSI.However,due to the impact of channel estimation errors,quantization errors and feedback errors,it is impossible to achieve perfect CSI.In order to overcome the CSI uncertainties,in this thesis,a heuristic robust relay selection algorithm and an optimal robust power allocation algorithm are proposed in cognitive relay networks to maximize the capacity of cognitive network under the total power budget constraint of SU-TX and relay nodes.Based on the Worst-case robust optimization theory and Lagrange dual decomposition method,the robust power allocation problem is solved and a closed form analytical solution to the robust power allocation algorithm is derived.Performance analysis validates the effectiveness of the proposed robust resource allocation algorithm.Simulation results show the proposed robust resource allocation algorithm is superior to the non-robust resource allocation algorithms proposed under perfect CSI in terms of guaranteeing the interference threshold of PU-RX,but at the expense of capacity loss of cognitive network.As for the forementioned robust resource allocation algorithm which is proposed under the condition of single PU-RX and individual power constraint,we make three improvements in this thesis as follows: 1)Joint control the transmit power of the SU-TX and the relay nodes in order to make the power allocation process more flexible.2)Deploy multiple PU-RXs with different interference threshold requirements in the primary network,which is more suitable to practical wireless communication scenarios.3)An optimal robust relay selection algorithm is proposed to make the robust power allocation algorithm achieve better performance.According to the above improvements,a robust resource allocation algorithm is proposed under the condition of multiple PU-RXs and joint power control constraint.This algorithm can maximize the capacity of cognitive network while maintaining the interference thresholds of multiple PU-RXs.Additionally,traditional resource allocation algorithms with QoS optimization targets can not reflect users' subjective satisfaction and will cause a waste of radio resource.Currently,in order to directly indicate subjective opinions of users and improve spectrum efficiency,most of the existing radio resource algorithms adopt QoE as the optimization target.However,guaranteeing users' QoE will consume large amounts of energy.Motivated by the aforementioned discussions,this thesis studies the resource allocation problem to jointly optimize SUs' QoE and energy consumption in two-way cognitive relay networks.The overall SUs' QoE per power consumption is defined as quality of experience per Watt(QoEW),which is regarded as the trade-off metric to improve overall SUs' QoE and decrease power consumption.In this thesis,a resource allocation algorithm based on QoEW is proposed in two-way cognitive relay networks.The optimization objective is to maximize the QoEW under the constraint of maximum total transmit power of SUs and relay nodes,while guaranteeing the minimum QoE requirements of SUs and keeping the interference power to multiple PU-RXs below their tolerable thresholds.Based on cross-layer optimization architecture and Lagrange dual decomposition method,the optimal QoEW is achieved.Numerical simulation results show the effectiveness and superiority of the proposed algorithm,and demonstrate the impact of system parameters on the QoEW.Finally,due to the topology structure of cognitive relay networks is very complex,it is difficult to implement the procedure of encryption key distribution and encryption key management with traditional encryption method in cognitive relay networks.Physical layer security technology can fully exploit random characteristics of wireless channel to ensure the information safety.It does not need to generate encryption key and has the advantage of low complexity.Therefore,physical layer security technology is suitable to ensure the information safety of cognitive relay networks.There is also energy consumption problem when ensuring the physical layer security of cognitive relay networks.Motivated by the aforementioned discussions,this thesis studies the resource allocation problem to jointly optimize physical layer security and energy consumption in cognitive relay networks with an eavesdropper.The ratio of the secrecy rate of SU-RX to power consumption is defined as secrecy rate per Watt(SRW),which is regarded as the trade-off metric to increase the secrecy rate of SU-RX and reduce power consumption.In this thesis,a resource allocation algorithm based on SRW is proposed in cognitive relay networks.The optimal objective is to maximize the SRW under the constraint of maximum total transmit power of SU-TX and relay nodes,while guaranteeing the minimum secrecy rate requirement of SU-RX and keeping the interference power to each PU-RX below its tolerable threshold.Based on Dinkelbach method,Lagrange dual decomposition method and difference of convex(DC)programming,the optimal SRW is achieved.Numerical simulation results show the proposed algorithm can ensure the minimum secrecy rate of SU and improve the SRW of cognitive network effectively.