Detection of narrow-band sonar signals using order statistical filters

In detecting a narrow-band signal in noise in a sonar environment, an estimation of the background noise power has to be made so that a detection threshold can be set. Here, we consider a variety of nonlinear estimators based on order statistics. The performances of these methods in providing an estimate of the noise power and in detecting a narrow-band signal in noise have been analyzed and have been found to be robust even under the interference of other signals.