Adaptive Access For ISM Band Cogntive Radio

Cognitive Radio(CR)technology can solve the problem of unbalanced allocation of spectrum resource properly,which can help cognitive users opportunistically access the idle frequency bands on the basis of perceivsing the shared spectrum,thereby the contradiction between resource shortage and low resource utilization will be settled effectively.Spectrum prediction can help cognitive users to make rational use of frequency selection by learning and grasping historical spectrum information,and to reduce the expended time and energy of spectrum sensing process.Reasonable allocation of cognitive resources can increase the utilization rate of precious spectrum resources and improve spectral efficiency of cognitive systems.This article focuses on spectrum prediction technology and cognitive access optimization technology.The main contrabutions are as follows:First,by studying of spectrum prediction techniques in cognitive communication scenarios deeply,the principle,application scenarios and performance of the existing spectrum prediction algorithms analyzed are.A spectrum prediction algorithm based on long and short term memory network(LSTM)model is proposed,it has excellent performance when dealing with the data with time series features.This paper has carried on the reasonable structure design and the necessary optimization work to the LSTM algorithm.In particular,based on the prediction and inference of the time series of 0-1 occupancy,the LSTM algorithm is also used to predict the power level of the spectrum.Second,the problem of resource allocation of downlink in cognitive satellite scenario is analyzed in detail.Based on the actual situation of cognitive terminal in the scenario,a downlink resource allocation model of cognitive satellite is established.In this model,cognitive terminals are equipped with spectrum sensing modles,and the preferred frequency channels of the cognitive terminals are not the same.The satellite combines the demand and frequency channels of cognitive terminals to make timely and efficient resource allocation decisions.The model is designed to maximize the throughput and meet the throughput requirements of cognitive terminals,it includes three-dimensional resources,namely,payload power,time and dedicated spot beams.By using the Lagrange duality algorithm,the model can be solved optimally.Third,by validating with actual data and simulating with a large number of reasonable tests,respectively,the superiority of the LSTM model and cognitive satellite resource allocation algorithm is proved.In particular,the LSTM network algorithm is superior to the previously proposed HMM,NN and other algorithms,it also shows better performance in predicting the power sequence.The proposed resource allocation algorithm can realize the distribution of the three-dimensional resources rationally,that is,the throughput of the system is maximized on the basis of that the throughout requirements of the cognitive terminals are satisfied.Thus a reliable premise is established for the fast and efficient cognitive access process.