Performance evaluation of eigenvalue-based detection strategies in a sensor network
Eric Ayeh, Kamesh R. Namuduri, Xinrong Li · 2014
The detection of random signals in noisy measurements is a problem of interest in several scientific applications that has been studied extensively. Recently, sample eigenvalue-based procedures for spectrum sensing and signal detection have received a lot of attention due their computational simplicity, their robustness, and their performance, which is claimed to exceed the performance of the classical Neyman-Pearson detectors. In this paper, through a theoretical analysis and Monte Carlo simulations, we investigate the detection performance of the different eigenvalue-based detection strategies that have been proposed while utilizing the performances of the energy detector and the estimator-correlator as benchmarks. Our results indicate that eigenvalue-based methods are not better than the classical energy detector and estimator-correlator.