Optimal Spectrum Sensing in Cognitive Radio Systems Using Signal Segmentation Algorithm

In the last two decades, wireless spectrum resource shortage has been the bottleneck in the development of wireless communication systems. More than that, many measurement campaigns confirmed the fact that the frequency spectrum is inefficiently used. Cognitive radio is a promising technology for the efficient allocation and usage of the available spectrum. It identifies the available spectrum and reallocates them for efficient utilization of frequency bands. In this paper, a novel spectrum sensing technique is proposed. The proposed method involves two levels of segmentation in time domain. In first level, running mean of the samples are considered and compared with the threshold. In second level, difference statistics is taken and the segmentation of signal is carried out. Further, frequency of the detected signal is computed, and the graph between the probability of detection and Signal to Noise Ratio (SNR) is plotted for input signals of various SNR. Two previous methods were considered for performance comparison, they are energy detection method, covariance based spectrum sensing method. The proposed method performs better than the above mentioned methods, especially in low SNR values.