Research On Efficient Spectrum Allocation In Cognitive Networks

Prediction and allocation techniques based on spectrum hole are the key technologies to solve the problem of low spectrum utilization.Starting from the prediction of free spectrum distribution and the efficient spectrum allocation strategies,this thesis conducts research on more accurate prediction mechanism and more efficient spectrum allocation strategies.The main research contents and innovation work are summarized as follows:1.The current spectrum prediction model assume that all the users have the same demand and occupy the single channel to transmit information,which is inconsistent with the actual service requirements of users and cannot accurately predict idle spectrum.Aiming at this problem,this thesis proposes a spectrum prediction mechanism under the condition that user demand changes and the amount of occupied spectrum is uncertain.According to the continuity of the spectrum occupied by user,this mechanism transforms the user with high demand and multi-channel transmission into single channel transmission,the maximum continuous free spectrum is used as the prediction object,and its probability distribution is predicted by the idea of mathematical induction,and the prediction result is used as the basis for the single cognitive user to access network.At the same time,the prediction mechanism of available free spectrum under the condition of multiple cognitive users is established.The simulation results show that the newly established model is more accurate than the existing model in free spectrum prediction,and the free spectrum prediction mechanism is reasonable for multi-user condition,which can effectively grasp the total amount of free spectrum in the network.2.The existing spectrum allocation strategies ignore the accuracy in spectrum sensing and cause quality problems for cognitive users to apply for free spectrum,users cannot complete data transmission through free spectrum with poor quality,and fairness is difficult to guarantee.In order to solve this problem,this thesis proposes a spectrumallocation algorithm based on spectrum quality and reward mechanism.The algorithm adopts the spectrum credibility index to characterize the quality of the free spectrum,and introduces spectrum credibility index into the construction of the price function to distinguish cognitive users with different detection capabilities by adjusting the unit price of the spectrum.Moreover,introducing reward mechanisms into utility functions,cognitive users with strong detection ability and high credibility index have high rewards,while cognitive users with poor detection ability and low credibility index have low rewards,so that the cognitive users with strong detection ability are able to apply more free spectrum,and the utilization of the free spectrum is improved.The simulation results show that the newly developed spectrum allocation strategy can distinguish cognitive users more effectively,and carry out survival of the fittest for cognitive users,so as to improve the number of cognitive users with high detection ability and improve the overall spectrum utilization of the network.