The Study Of Optimization Of Spectrum Sensing In Cognitive Radio

Cognitive radio (CR) has recently been introduced to solve the problem of low spectrum utilization through opportunistic spectrum sharing. Reliable and high-quality spectrum detection technology is an important prerequisite that improves spectrum efficiency and protects licensed users. MAC layer spectrum sensing techniques in CR have been studied based on physical layer detection algorithm in this thesis, which includes: joint optimization of detection period and detection time, optimal algorithm for searching idle channels and channel selection.Under single-user detecting multi-channel scenarios, this thesis studies the joint optimization method of detection period and detection time to improve spectrum efficiency and normalized throughput of CR network, while mitigating the interference to licensed users. Firstly, two factors are introduced: the discovered spectrum access opportunities ratio and the interference to licensed users. The optimization function of detection period is proposed to maximize spectrum efficiency and decrease the interference to licensed users, which has been conformed by simulation. Secondly, under the target detection probability constraints, detection time optimization problem is studied to maximize the achievable normalized throughput of CR network. At the last, this thesis develops the joint-optimization algorithm of detection period and detection time based on single-optimization methods of detection period and detection time. From the simulation, It can be found that there does exist an optimal detection period, which greatly reduces the interference to licensed users with only slightly decreases the spectrum access opportunities ratio. Compared with single-optimization methods of detection period and detection time, the joint optimization algorithm can greatly improve the normalized throughput of CR network.Under single-user detecting multi-channel scenarios, CR user must immediately exit the channel and search other idle channels when licensed users reoccupy it. Considering idle probability and utilization probability of a channel, this thesis develops an optimization algorithm to search an idle channel. This algorithm introduces two factors: idle probability and utilization probability of a channel, which are combined to a selection function of an idle channel. This thesis derives the expression of average delay that CR user searches the first idle channel and discusses the impact of waiting time to average delay. It is observed from the simulation that simulated and theoretical curves exhibit the same trend and their values are almost the same, which verify the correctness of theoretical analysis in this thesis. Our algorithm can greatly reduce the searching delay compared with random searching method.Under multi-user detecting multi-channel scenarios, CR users can choose different channels to perform spectrum sensing, this thesis studies the channel selection problem based on spectrum sensing. Firstly, under the limit of target detection probability, detection time optimization method is developed to get the maximal normalized throughputs of different CR users detecting different channels. Then, the improved Hungarian algorithm is proposed to select the optimal channels for CR users. Simulation results show that the improved Hungarian algorithm can greatly improve the overall normalized throughput comparing with the random channel selection method.