Design of an Improved Spectrum Sensing Technique Using Dynamic Double Thresholds for Cognitive Radio Networks

Abstract

We, in this paper, introduce a spectrum sensing scheme utilizing two stages for spectrum detection and utilizing dynamic threshold. The first stage consisting of energy detection is followed by fuzzy logic based detector employing dynamic thresholds for final decision making of the spectrum sensing. Two metrics—Credibility_energy and Credibility_psd as inputs to the fuzzy logic detector are introduced. Numerical results show that the proposed scheme improves detection performance significantly as compared to existing two stage schemes at local node. The approach does not require a priori knowledge about the format of primary user signal retaining its practical implementation simple. We further investigate cooperative sensing with the proposed scheme to show that it outperforms the conventional cooperative spectrum sensing. An analytical expression for optimal number of CR users for cooperative spectrum sensing is also derived. This yields enhanced local sensing performance with reduced communication overheads. Having determined the optimal number of users, for cooperative spectrum sensing, the probability of error defined as the sum of probability of miss detection and probability of false is also minimized. Numerical results show that more enhanced local sensing performance reduces the number of optimal users for cooperation up to a significant extent at low SNR. Owing to the advantages of easy implementation, low cost and smaller overheads, the proposed solution offers a potential spectrum detection technique for cognitive radio systems.