An adaptive threshold method for energy based spectrum sensing in Cognitive Radio Networks

Today, communication systems face major shortages in availability of free spectrum. Cognitive radio has evolved as the most feasible technology which can empower us to effectively utilize the available spectrum. A vital component of a Cognitive Radio Network is spectrum sensing, i.e. its ability to detect the presence of a primary user. This paper employs an adaptive threshold detection algorithm based on an image binarization technique. A novel approach is proposed to fix the standard deviation threshold whereby the system dynamically trains itself based on previously iterated decision statistics. Such data is computed on the basis of calculated probability of detection, probability of false alarm, standard deviation coefficient and ratio of occupied bandwidth to the total bandwidth. Simulation results show the vastly improved performance compared to the conventional energy detector. At low Sound to Noise Ratio the performance improvement index shows an increase of 30 percent in comparison to the conventional energy detector.