Research On Resource Management In Energy Harvesting Cognitive Radio Networks

With the widespread popularity of new technologies such as industrial internet of things and mobile internet industry,wireless devices and related applications have increased dramatically.The contradiction between the huge demands for wireless spectrum and the limitation of spectrum resource becomes more prominent.Cognitive radio is emerging as a promising technology to achieve the dynamic spectrum access,which could improve the spectrum efficiency so as to fulfill the growing spectrum requirements.In cognitive radio networks(CRNs),secondary users(SUs)are available to use the authorized spectrum for communication without harming primary users'communications,so as to improve spectrum efficiency.Aiming to fully exploit the idle spectrum,SUs need to perform spectrum sensing,spectrum analysis,as well as spectrum handoffs.These operations along with the related data transmissions result in high energy consumption.Thus,how to effectively improve energy efficiency is one of the issues that need to be solved urgently in CRNs.Energy harvesting(EH),which is used to replenish energy from ambient energy sources to power wireless equipements,has the advantages of no pollution and unlimited potential available energy.Therefore,EH has been flagged as one of the effective approaches for improving energy efficiency.In light of the above advanced features,,applying EH in CRNs give the birth of EH CRNs,which is able to improve both spectrum and energy efficiency and has become increasingly eye-catching recently.Although some progress has been made in the energy harvesting resource management schemes,most of them mainly focus on the traditional communication networks.Since cognitive radio networks have unique functions such as spectrum sensing,spectrum analysis and spectrum handoffs,existing energy harvesting resource management schemes have poor performance or no longer applicable in cognitive radio networks.Besides,existing energy harvesting resource management schemes mainly consider the static networks.There is less research on the planning of resource in CRNs whose system state(including channel state,energy harvesting state,etc.)dynamically changes.The thesis focuses on resource management problem in EH CRNs and tackles the aforementioned issues as follows:1.In EH CRNs where energy harvesting and the data transmission perform sequentially,we study the joint optimization of energy harvesting duration and transmission power allocation.We propose an optimal policy which is able to dynamically adjusting the energy harvesting duration as well as the transmit power according to the system state.In addition,we introduce a sub-optimal policy,and the closed-form expression between the sub-optimal policy and the network parameters is deduced.Specifically,1)considering the diversity of channel states,the availability of authorized spectrum,and the existence of sensing errors,we formalize the evolvement of system states,and construct the value function which represents the long-term throughput;2)we propose an optimal policy,which is able to specify the energy harvesting duration and the transmit power according to the current system state;3)it turns out that there exists a threshold determined by the battery storage capacity,and the optimal long-term throughput is monotonically increasing with the amount of available energy in the battery in case that the available energy is under the threshold;4)we introduce a sub-optimal policy,and theoretically prove the closed-form expression between the sub-optimal policy and the network parameters,thereby greatly reducing the computational cost for deriving the sub-optimal policy.2.In EH CRNs which perform periodic spectrum sensing,we recognize that the spectrum sesing improves the accuracy of the sensing results at the cost of consuming both time and the energy resource.Besides,a reasonable transmit power allocation is very important for enhancing the system performance.We propose the optimal policy which jointly optimize the sensing duration and the transmit power allocation,so as to minimize the long-term outage probability.Besides,a low-complexity sub-optimal policy is also introduced.Specifically,1)considering the channel states,available energy and the energy replenishment process,we formalize the evolvement of the system states,and propose the value function which represents the long-term outage probility;2)aimming to minimize the long-term outage probability,we propose an optimal policy which is able to adjusting the sensing duration as well as transmit power based on the current system state.We prove that the optimal long-term outage probability is monotonically decreasing with the amount of the available energy;3)we introduce a sub-optimal policy with reduced computational complexity.As the Signal to Noise Ratio(SNR)of the received signal increases,the performance of the sub-optimal policy is approaching to the optimal policy,and eventually achieves the same performance.3.In EH CRNs which perform periodic spectrum sensing,we analyze the problem of maximizing long-term throughput through jointly optimizing channel sensing duration and transmit power.We propose an optimal policy for maximizing the long-term throughput.It is proved that if the received primary SNR is sufficiently high,the optimal transmit power is monotonically increasing with the available energy.In addition,we introduce a sub-optimal policy with low-complexity.Specifically,1)we formalize the evolvement of the dynamic system states,and design the value function which represents the long-term throughput;2)we propose an optimal policy which is adaptive to the system states by dynamically changing sensing duration as well as transmit power,so as to maximize the long-term throughput;3)it is theoretically proved that when the received primary SNR is sufficiently high,the optimal transmit power is monotonically increasing with the available energy.A sub-optimal policy with low-complexity is introduced.The performance of this sub-optimal policy is comparable to that of the optimal policy when received primary SNR is sufficiently high.4.In EH CRNs which carry out spectrum sensing on demand,we investigate the problem of resource optimization in terms of spectrum sensing decisions as well as dynamic allocation of transmit power.We propose an optimal resource management policy which dynamiclly makes spectrum sensing decisions and determines the corresponding transmit power.In addition,we introduce a sub-optimal policy,and prove that the sub-optimal policy has a threshold structure between the spectrum sensing decision and the available energy.Specifically,1)allowing for the system state comprised of channel condition,energy harvesting process,and the available energy,we derive the state transition probability distribution corresponding to different spectrum sensing decisions as well as transmit power,and formalize the evolvement of the system states.A value function which represents the long-term throughput is introduced;2)we propose the optimal spectrum access policy.The optimal policy specifies the spectrum sensing decision as well as the transmit power according to the current system states;3)we design a sub-optimal policy,and prove that there exists a threshold structure between the spectrum sensing decision and the available energy.An efficient algorithm is proposed to obtain the sub-optimal policy.