Performance evaluation of Welch's periodogram-based energy detection for spectrum sensing

In this work, the authors present the evaluation of energy detection (ED) based on the Welch's periodogram for spectrum sensing applied to cognitive radio networks. The authors analyse the impact of the number of points in the discrete Fourier transform and the number of averaged periodograms for power density spectrum estimation on the performance of ED. The authors identify that the inclusion of these parameters in the distribution of the test statistic used to detect the presence of primary users, improves the probability of detection. However, in the presence of noise uncertainty, the improvement on the probability of detection will come at the expense of an increased probability of false alarm. With the approach considered in this work is possible to increment the probability of detection for a given and low signal-to-noise ratio, without increasing the number of samples collected from primary signal. However, to maintain a constant probability of false alarm, accurate techniques for noise variance estimation are needed, because detection-threshold value is highly dependent on the noise power present at each sensing interval.