Research On The Spectrum Decision Based On Spectrum Prediction And Spectrum Recommendation In Cognitive Radio

In recent years,due to the rapid development of wireless communication,the number of wireless communication equipment and the demand of wireless communication increase gradually,the spectrum resources become more precious and scare.At present,the spectrum allocation method mainly takes the fixed allocation strategy,which has caused the waste of the spectrum resources.Due to the demand of development,cognitive radio technology emerged.With the intelligence of cognitive abilities,it can find out the spectrum hole and make use of the spectrum hole through spectrum sensing,dynamic spectrum access technology and so on.The purpose of cognitive radio is to improve the utilization of spectrum.In this paper,we focus on the key technology of cognitive radio,spectrum forecasting and spectrum decision technology,as well as the new proposed spectrum recommendation.We propose a spectrum decision framework model based on spectrum prediction and spectrum recommendation,and then study on each part separately.Firstly,the extreme learning machine method is applied in the spectrum prediction technology,and the spectrum prediction algorithm based on ELM is proposed to predict the spectrum state of future.Compared with the traditional neural network method,the speed of the training is greatly improved and the performance of the forecast is also ensured.In this paper,a cooperative spectrum predictin based on Q-learning algorithm is also proposed.Compared with the local prediction of single user and the traditional cooperative prediction method,the busy or idle state of the bands can be predicted better.Then,the spectrum recommendation method based on the cooperative filtering is applied to the spectrum decision frame in this paper.By the experimental verification,the collision rate between users is obviously reduced,and the utilization of the channel is improved by using spectrum recommendation.Finally,we combine the spectrum prediction with spectrum recommendation by Q-leaming algorithm and Markov decision process(MDP)to realize pectrum decision.According to the user location and communication range,the system scene is divided into two scenarios,and we finish the experimental simulation respectively.The experimental results show that the spectrum decision method based on spectrum prediction and spectrum recommendation can reduce the collision rate between users and improve the efficiency of spectrum utilization.Spectrum decision determines whether the cognitive radio system can make more use of the spectrum resource in the case of the low user conflict rate.The spectrum decision method based on spectrum prediction and spectrum recommendation is a new spectrum decision idea,and it is meaningful to the further research of spectrum decision technology.Spectrum decision technology has promising application prospects in the intelligent WLAN and the military anti-jamming.The effective spectrum decision method can improve the utilization of the spectrum,and make the most of the spectrum resource in the fields such as phone communications,military,spaceflight,and so on.It is meaning to the development of social and economy.