On Optimization Of Spectrum Sensing For Cognitive Radio With Energy Harvesting

With the advancement of wireless communications,more and more users require to access to the wireless spectrum resource.However,the shortage of wireless spectrum resource has impeded these requirements.In the meanwhile,the allocated spectrum is seriously underutilized.Hence,it is important to improve the utilization efficient of wireless spectrum resource.Cognitive radio,as one of cutting-edge technologies,is a promising solver for the problem.The growing number of wireless users implies higher energy consumption at the user ends.In some scenarios,it is impossible to maintain the power supply for user ends powered by batteries with limited amount of energy.Energy harvesting makes it possible through self-sustainable in power supply.In this thesis,we combine cognitive radio and energy harvesting to improve both wireless spectrum utilization and energy efficiency,and then achieve a maximum throughput.The main contributions are outlined as follows.(1)In the scenario of an energy harvesting secondary(SU)synchronized with primary user(PU),the length of a SU frame is predefined,consisting of three separate phases – energy harvesting,spectrum sensing and data transmission.In order to maximize the throughput of SU,we optimize the intervals of the three phases.Within the optimization,we first eliminate the effect of a random variable by manipulation and transformation,and then we propose a greedy iteration algorithm which reduces the computational overhead.(2)In the scenario of an energy harvesting SU unsynchronized with PU,the length of a SU needs to be optimized with the states of PU considered to follow the Possion distribution.In order to maximize the throughput of SU,we develop the problem as a joint optimization in relation with the length of a frame and that of the energy harvesting.To reduce the computational complexity,a greedy alternating-direction iteration method is applied in the optimization,where we furthermore derive the throughput upper limit.(3)In the scenario of an energy harvesting SU synchronized with PU,considering the case that the length of spectrum sensing is negligible in comparison with the length of a frame,we assume that the energy consumed in the spectrum sensing can be ignored.Meanwhile,we use the “energy packet” to model the energy harvesting and energy consumption process,and a “simultaneous harvesting-consuming” protocol is applied,where the energy packets can be harvested or consumed simultaneously.In order to maximize the throughput of SU,we develop the objective function with respect to energy threshold to be optimized.