Optimum Detectors of Line Spectrum Signals in Generalized Gaussian Noise

Abstract

Optimum signal detectors are investigated based on the underwater ambient noise collected in South China Sea in the winter of 2006. The statistics of underwater ambient noise has been firstly analyzed, and the results lead to a conclusion that the generalized Gaussian process is very suitable for ambient noise statistically modeling. Thereafter, the optimum detectors have been deduced for line spectrum signal detection, and the asymptotic forms of these optimum detectors have also been derived under the assumption of weak signal. It has been further proved that the optimum detectors in Gaussian noise are the specific cases of those in the generalized Gaussian noise.