On Spectrum Sensing And Dynamic Resource Allocation For Cognitive Radio Systems

The radio spectrum is a valuable non-renewable resource in communications.With the rapid development of the wireless communication technique and the increas-ing user’s needs of diverse data services, the contradiction between the scarce spec-trum resource and the user’s needs becomes more and more serious. However, thesurvey indicates that the traditional fixed spectrum assignment policy as well as theimbalanced use of the allocated spectrum is the nature of this contradiction. Cognitiveradio is considered to be a promising method which effectively employs spare spec-trum and solves the insufficiency of the wireless spectrum. It can intelligently sensethe surrounding radio environment and automatically searches for the available spec-trum, dynamically reassign and reuse this spectrum. Therefore, the use of cognitiveradio opens up a new way to improve the spectrum utilization.Due to the random activities of licensed users, also called primary users, theavailable spectrum is also random. How to quickly and efficiently obtain a reliableinformation on the spectrum usage is the cognitive user’s primary task, as well as thegoal of spectrum sensing. Based on the sensing results, the cognitive radio user, alsocalled secondary user or unlicensed user, will adjust its operation parameters to adaptto the changes of the surrounding environment, while protecting the licensed users’transmissions from interfered with. How to manage the interference to the licensedusers and the mutual interference among cognitive radio users to achieve the optimalperformance is the main task of dynamic resource allocation among cognitive radiousers due to the characteristics of cognitive radio systems, which should be carefullydesigned. Therefore, this dissertation makes in-depth study on spectrum sensing anddynamic resource allocation of the cognitive radio systems. The main findings are asfollows: Firstly, the optimization of broadband spectrum sensing in cognitive radio sys-tems is studied based on energy detection. In an OFDM-based broadband system,each subcarrier experiences different fading due to time-varying channels and multi-path fading. How to quickly find and identify the weak licensed user’s signal requiresdifferent decision thresholds to be set on each subcarrier. Besides, the sensing timeaffects the effectiveness and accuracy of the sensing result, which should be consid-ered together with the decision threshold. Therefore, a spectrum sensing problem inwhich the sensing time and the decision threshold are optimized jointly is consideredfor OFDM-based broadband systems to achieve the optimal performance of cognitiveradio systems with constraints on the sensing accuracy requirement and the interfer-ence to the licensed user. The problem is solved using optimization theory and a jointoptimization algorithm is proposed. Numerical results show the proposed algorithmcould maximize the opportunistic throughput of cognitive radio systems and greatlysatisfy the requirements.Secondly, the sensing time is limited by the traditional frame structure of cogni-tive radio systems, resulting in an inherent tradeoff between the sensing time and thesystem throughput. To overcome this inherent tradeoff, the structure of the cognitiveradio user’s receiver is improved at first, in which the demodulation of the receivedsignal and the spectrum sensing are jointly designed to allow data transmission to-gether with spectrum sensing. Accordingly, the frame structure is improved as well.To reduce power consumption, a parallel spectrum sensing method with sleep modeis proposed, in which the sensing module in the improved receiver would be switchedinto sleep mode to save power when the sensing accuracy requirement is met. Con-sidering licensed users’ activities, a joint optimization problem of sensing time andpower allocation is studied which maximizes the achievable opportunistic through-put of cognitive radio systems while satisfying the sensing accuracy requirement andpower budget and keeping the interference to licensed users below a given thresholdlevel. A joint optimization algorithm is proposed in which the optimal power alloca-tion is found at a given sensing time using an inner iteration algorithm and the optimalsensing time and power allocation is obtained by a outer searching algorithm. Finally,numerical comparisons verify the feasibility and show the advantages of the proposedalgorithm by comparison with other methods. Thirdly, the existing dynamic resource allocation algorithms for cognitive radiosystems are studied based on OFDM and proposed in a static environment. The cog-nitive radio user’s moving or the other objects moving around a transceiver will leadto inter-carrier interference (ICI). Such ICI severely limits the system performanceespecially in high mobility environment. However, the most resource allocation algo-rithms proposed for OFDM-based cognitive radio systems neglect this ICI. Therefore,a simple adaptive subcarrier bandwidth method is proposed to reduce the impact ofICI on the system performance. On this basis, a joint optimization problem of sub-carrier bandwidth and power allocation in OFDM-based cognitive radio systems isfurther studied in which optimizing the subcarrier bandwidth and power allocation tomaximize the system bandwidth efficiency under constraints on the subcarrier band-width, power budget and the interference to the licensed user. An exhaustive iterationalgorithm is proposed. Numerical simulation results show that the proposed algorithmcould maximize the system bandwidth efficiency and balance this tradeoff while sat-isfying the constraints.Fourthly, it is found that the most dynamic resource allocation for OFDM-basedcognitive radio systems are studied to achieve the spectrum efficiency maximization.However, energy consumption and environmental protection have become global de-mands and inevitable trends today. Considering the life of battery power and networksustainable development, minimizing the energy consumption of cognitive radio com-munications not only reduces environmental impact, but also cuts overall networkcosts and helps make communication more practical. Therefore, under opportunis-tic spectrum access, dynamic resource allocation problem in multiuser OFDM-basedcognitive radio systems with real-time applications is studied to minimize the totaltransmission power of cognitive radio systems while meeting the target transmissionrate requirements of cognitive radio users and the interference power constraint. Theoptimal subcarrier assignment and power allocation are analytically derived and a jointoptimal allocation algorithm is proposed. Due to its high complexity, a suboptimalalgorithm based on stochastic nature is proposed. Numerical results show that theperformance of the suboptimal algorithm with low complexity is close to that of theoptimal one, which demonstrate that the suboptimal algorithm outperforms the otheralgorithms. Fifthly, due to the lack of cooperation between licensed users and cognitive ra-dio users in non-cooperative spectrum sharing, less spectrum chances are available forcognitive radio users. On the other hand, the side information about the licensed userthat is obtained by cognitive radio users is uncertain due to the asymmetry of the spec-trum access rights, which affects the choice of precoding strategy at transmitter. Inaddition, considering the interference power constraint, the performance of cognitiveradio systems has been greatly restricted. To this end, a cooperative spectrum sharingmethod is investigated for the asymmetry information case in which the cognitive radiouser acts as a cooperator of the licensed user to assist the licensed user’s transmissionand is allowed to simultaneously transmit its own data at the same spectrum. Ex-ploiting the difference in transmission performance, the cognitive radio user employssuperposition coding techniques to eliminate the interference from the licensed user.Furthermore, a power allocation problem in OFDM-based cognitive radio systems un-der cooperative spectrum sharing with real-time applications is studied to minimize thetransmission power of cognitive radio systems with constraints on the cognitive radiouser’s target rate requirement while guaranteeing the licensed user’s instantaneous rateunchanged if the cognitive radio user allocates part of its power for its own transmis-sions. This two-variable constrained optimization problem is transformed into a singlevariable constrained optimization problem and a greedy-based power allocation algo-rithm is proposed. Numerical simulations show that the proposed algorithm greatlyoutperforms other methods in the same environment.Finally, a conclusion is drawn for the dissertation, and valuable research direc-tions in the future are discussed.