Performance Analysis of Spectrum Sensing in Cognitive Radio Networks Based on AD Goodness of Fit Under Noise Uncertainty

Abstract

AbstractSpectrum sensing is one of key technologies for primary user detection in cognitive radio networks, and Goodness of Fit Test exhibits good sensing performance. In this paper, we analyze the performance of spectrum sensing based on AD goodness of fit (AD test) under noise uncertainty, and compare it with traditional energy detection. Moreover, much attention is paid on the sensing performance of AD test and energy detection under lower SNR, and analyzes the impact of direct current (DC) component for AD test when the noise power is uncertainty. We first give formulas of AD test and signal expressions with different percentage of direct current (DC) component, and discuss the sensing performance by using Monte Carlo simulation in AWGN channel and Rayleigh channel. Then the impact of noise uncertainty on spectrum sensing performance of AD test is discussed and compared with energy detection in different network environments. Extensive simulations and theoretical analysis indicate that spectrum sensing performance of AD test is better than energy detection when SNR is lower and there is no DC component in the transmitted signals under noise uncertainty. The above conclusion shows that AD test is less sensitive to noise uncertainty traditional energy detection.KeywordsCognitive radioSpectrum sensingAnderson-Darling testEnergy detectionNoise uncertainty