Resource Allocation Strategies And Performance Analyses In Cognitive Radio Networks

With the rapid development of wireless communications and the explosive growth of new wireless networks and services,the scarcity of spectrum resource has become a serious problem.However,the major reason of the shortage problem is that the low utilization efficiency brings about a waste of the spectrum.Cognitive radio network(CRN),as a spectrum sharing based technology,is able to improve the utilization efficiency of the licensed spectrum.By means of reasonable resource allocation(RA)strategies,the secondary users(SUs)can access the licensed spectrum opportunisticly and guarantee satisfactory communications for the primary users(PUs)simultaneously.Therefore,the researches on the RA strategies and their performance analyses are very important for the CRNs in different situations.In this dissertation,innovative RA strategies and the corresponding performance analyses are provided in four kinds of CRNs.Firstly,in OFDMA based CRNs,a RA strategy is proposed based on the interference minimization.Secondly,in mobile CRNs,an anti-shadowing RA strategy is proposed.Then,based on the previous work,a RA strategy is proposed to achieve an energy efficiency(EE)and spectrum efficiency(SE)tradeoff between the PUs and the SUs.At last,in imperfect NOMA based heterogeneous internet of things(IoT),an energy efficient RA strategy is proposed.The main contributions of this dissertation are as follows.1.Aiming at the problem that the performance of the PUs in OFDMA based CRNs is limited by the SUs,a RA strategy is proposed to minimize the interference power from the SUs,which guarantees the demands of all the users and effectively improves the capacity of the PUs.At first,through assigning subchannels,allocating power and controlling spectrum access reasonably,the SUs and the PUs can share the licensed spectrums with the maximized primary capacity.Then,compared with the traditional strategies,the simulation results show that the proposed strategy can reserve the primary capacity better.Meanwhile,the Quality of Service(QoS)of the users and the power limitation of the secondary base station can be guaranteed.2.Aiming at the problem that the instantaneous sensing results of channel state information(CSI)cannot be directly used to allocate resources in mobile CRNs,an anti-shadow fading RA strategy is proposed based on CSI prediction,which effectively improves the EE and the SE of the mobile SUs.Firstly,the channel state information prediction based system model is considered.According to the prediction result,protecting parameters are preset to minimize the prediction error caused by the shadowing.Secondly,a RA strategy is proposed to maximize the average effective capacity of the SUs.Then,the speeds of the PUs are taken into account,and a general system model is built.Finally,simulation results verify that the proposed strategy can not only improve the total secondary capacity but also achieve higher EE and SE for the mobile SUs.3.Aiming at the difficulty of jointly optimizing the EE and the SE for both the PUs and the SUs in CRNs,a deep learning based RA strategy is proposed to minimize the weighted sum of the interference power from the SUs,of which the computational efficiency is obviously better than those of the traditional strategies.Firstly,a RA problem is formulated to minimize the weighted sum of the secondary interference power,where the performance of both the PUs and the SUs are guaranteed by the constraints on signal to interference plus noise ratio(SINR),power consumption and data rate.Secondly,a deep learning based damped three-dimensional messagepassing algorithm(MPA)is introduced to solve the problem.Then,a suboptimal strategy is deduced based on a two-dimensional MPA with a higher computational efficiency.Finally,performance analyses are provided through simulation results,confirming the benefits of the proposed strategy on multiple effectiveness.4.Aiming at the problem of mutual interference between users and devices in imperfect NOMA heterogeneous IoT,a RA strategy based on optimal user paring and joint EE optimization is proposed,which effectively improves the access opportunities,EE and SE of the system.Firstly,based on the CRN theory,a stepwise strategy is proposed to allocated the downlink power and the subchannels for the hotspot users and the smart devices,through taking them as PUs and SUs.Secondly,a deep recurrent neural network is introduced to improve the computational efficiency of the proposed strategy.Then,a priorities and rate demands based user scheduling method is supplemented,to coordinate the access of the heterogeneous users.At last,the simulation results verify that the proposed strategy is able to improve the EE and the SE of the imperfect NOMA based heterogeneous IoT.