Congreso: 2007 IEEE workshop on statistical signal processing (SSP-2007). Madison, Wisconsin, USA. 26-29 de Agosto 2007

Autores: O.A. Yeste-Ojeda, J. Grajal

This work explores the capabilities of spectral correlation based detectors for partially or completely unknown signals corrupted by Gaussian noise. The main contribution is twofold: Firstly, the cyclostationary model is shown to be less suitable for detection purposes than the stationary model. Secondly, analytic approximations are derived for the detection and false alarm probabilities, under the assumption of long observation time.