Autonomous Global Threshold Adjustment Algorithm For Energy Detection In Self-Reconfigurable Cognitive Radio Systems

This paper suggests a novel autonomous global threshold adjustment algorithm for energy detection (ED) in self-reconfigurable cognitive radio (CR) systems. The proposed method uses the power spectral density (PSD) as an input and generates an L-bin histogram in the first instance. Instead of maximizing the between classes variance as in the classical method, the second step consists of dividing the histogram into two equal parts and finding the maximum difference between histogram bins. Finally, the threshold setting and the distinction between the presence of the signal in the spectral band and the unimodality of the distribution depend on the position of the previously mentioned maximum. The proposed method is validated and compared to the other SATA (self-adaptive threshold adjustment) algorithms on real-life signals. The obtained results are promising and meet the requirements of the IEEE 802.22 standard with a detection rate higher than 90 % and a probability of false alarm lower than 10 %.