A New Spectrum Sensing Strategy for Dynamic Primary Users in Cognitive Radio

In this letter, a new spectrum sensing strategy is proposed for primary users (PUs), which are dynamic whereby it is assumed that at least one and possibly multiple PU status changes occur in the channel during the sensing period. According to our proposed strategy, the last status change point of the PU is estimated first using maximum likelihood estimation (MLE). Then, any known spectrum sensing method can be applied to the observed samples between this last status change point and the end of the sensing period. By this way, degradation in probability of detection due to PU traffic is lessened. Dynamic programming (DP) is used to accelerate computation of the MLE. We also propose a new cumulative sum (CuSum) based weighting scheme. Simulations indicate that the proposed sensing strategy provides an improvement in terms of probability of detection for all the considered spectrum sensing methods. Furthermore, the proposed CuSum based weighting scheme achieves the best probability of detection when employed together with the last status change point estimation strategy.