New Approach for Predicting Channel Parameter in Cognitive Radio Network

摘要

One of the essential issues in cognitive radio network, is the sensing strategy for the secondary users. Because of the hardware limitation and energy constrains, a secondary user may not be able to sense all channels simultaneously. Cognitive Radio researchers overcome this problem by discovering new methods to estimate the idle/busy channel-usage patterns in order to accurately predict the channel availability with limited number of samples. In this paper, we study two important parameters and their impacts on increasing the prediction accuracy of the channel-usage patterns. These parameters are the spectrum sensing period and the number of samples that are used in channel parameters estimation. For this purpose an algorithm for Maximum Likelihood (ML) estimator has been built to predict channel occupancy for different patterns of channel utilization. To overcome the sensing overhead, the algorithm has been built to adaptively sense the spectrum upon the pattern of usage of the primary activity. The number of samples that is used in estimating the channel's parameters is also adaptively chosen to meet the accuracy requirement and the transition probability of the primary users' occupancy.