Study On Selection Strategy Of Cognitive Radio Spectrum Based On The Theory Of Prediction

As an important tool of easing the strain on the resources, cognitive radio can sense and make use of the free spectrum, which has very significant contribution to improve the spectrum utilization. Spectrum prediction of cognitive radio can improve spectrum efficiency; saving energy consumption awareness; reduce the business conflict between authorized users and cognitive users. This paper presents prediction based on neural network. Prediction based on neural network is an artificial intelligence approach with low complexity and high accuracy. Papers focus on two main parts, the first part is to predict cognitive radio spectrum by using neural network, the other part is to select the best free spectrum by analyzing the results of prediction.The main work in this paper can be summarized as follows:(1)First, this paper describes the background of the topics. Detailed introduce the basic principles and key technologies of cognitive radio.(2)Briefly introduced the prediction theory, autoregressive prediction algorithm and the neural network theory, prediction algorithm are put forward based on the theory of neural network, analyzed the method of neural network, the experiment was simulated in the time domain and multi-dimensional time domain respectively.(3)Applying of dynamic spectrum selection algorithm.First,analyzing the availability and stability of predicted free spectrum and choosing the spectrum with two high parameters by setting appropriate threshold.And then analyzing whether conflict exist authorized users and cognitive users.Finally,comparing and analyzing the performance of dynamic algorithm applied in this paper and the general randomly selected access algorithm by simulation.