Single-threshold serial algorithm for detecting signals with unknown amplitudes and durations

摘要

The threshold algorithm for detecting an arbitrary-shaped signal with unknown amplitude and duration, which is observed against the background of additive Gaussian white noise, is synthesized. The possibility of reducing the decision-making time compared with the maximum-likelihood detection algorithm is demonstrated. The synthesized algorithm is analyzed on the assumption of sufficiently large signal-to-noise ratios. Asymptotic expressions for the detection-error probabilities are obtained. The results are specified for a pulse with a linear time profile.