Research On Perceptual Control Method Of Opportunistic Spectrum Access In Cognitive Radio Networks

With the development of modern wireless communication technology,the shortage of resources is becoming more and more serious,which leads people to think about the way of spectrum allocation.In order to solve the shortage of spectrum resources,researchers have proposed cognitive radio technology.The access mode of Cognitive Radio(CR)is opportunistic spectrum access.The cognitive user perceives the dynamic spectrum environment through spectrum sensing technology,and the opportunity to share the licensed frequency band with the authorized user,that is,the primary user.Therefore,the improvement of spectrum sensing performance is crucial in cognitive radio.This thesis focuses on the research of spectrum sensing control method,and builds a multi-user multi-channel prediction model based on BP neural network.In the thesis,the research status of spectrum sensing is first reviewed.Several commonly used spectrum sensing techniques are introduced.Combined with the emphasis on spectrum prediction in recent years,the spectrum prediction methods commonly used in spectrum sensing are introduced.In order to realize the multi-user multi-channel network,the modeling and research of the single-user multi-channel model is completed first,and finally the multi-user multi-channel model is built.In the thesis.in the process of constructing single-user multi-channel model,the ON/OFF channel state modeling is selected.The theoretical analysis and simulation experiments are carried out on this model,and an important experimental parameter is obtained.The channel can be used for the remaining time.Combining the analysis of existing data in the channel(such as channel communication traffic occupancy,signal-to-interference ratio,channel idle utilization,user location,etc.),the access mode of the channel model is divided into three types.They are table access switching,distance access switching,and random access switching.From the theoretical analysis and experimental results,it can be seen that the number of switching according to the table switching control only accounts for about 10%of the switching times of the distance switching control,and the number of switching of the random access is also reduced to about 50%of the distance switching.However,since there is no calculation of the waiting time for the random switching,the random switching mode can be ignored.From the comparison of the other two switching modes,the reduction of the energy consumption can be clearly seen.In addition,in the construction of the single-user multi-channel model,the theoretical analysis and simulation experiments of the spectrum sensing period in the channel are also carried out.In IEEE802.22,the definition of the spectrum sensing period is to fix the sensing period.Here,the sensing period is set to dynamically change the time,and the optimal sensing period is obtained by minimizing the consumption function of the sensing period.In the four state models of the channel,the effective communication duration and the invalid communication duration in the four states are obtained by using the integral,and the probability of each state is used to update the probability value using the Bayesian derivation formula to form a channel model.In the channel model,the normalized channel efficiency and the normalized interference duration are taken to a minimum value to obtain an optimal data transmission time,and the optimal sensing time value can be obtained according to the data transmission time.The amount of data transmitted and the state of the channel itself are used to determine the sensing period of the spectrum.If the above two types of data change,the sensing period also changes.This quantifies the data selection and switching of channels in multiple-channels,and analyzes the excellent channels from the data to improve the data transmission rate and reduce channel interference.The single-user multi-channel model is restrictive in the number of cognitive users.In order to have a broader model and closer to the network mode actually used,a multi-user multi-channel model is built based on a single-user multi-channel model.In the multi-user multi-channel model,the BP neural network-based spectrum prediction is selected to make a short-term prediction of the remaining channel usable time,and the linear analysis is used to obtain the screening table and the intra-band spectrum sensing dynamic sensing period is used to obtain the perceptual period value to complete the network model.Through theoretical analysis and experiments,it is proved that the spectrum utilization increase percentage of the multi-user multi-channel system built is greater than 90%when the spectrum prediction plus channel control access method is used and the intra-channel perceptual period optimization is performed.The minimum percentage of perceived energy reduction is also greater than 46%.Therefore,the establishment of multi-user multi-channel prediction model based on BP neural network has practical significance.