Robust minimum distance Neyman-Pearson detection of a weak signal in non-Gaussian noise

In practice, noise distributions usually are not Gaussian and may vary in a wide range from light-tailed to heavy-tailed forms. To provide robust detection of a weak signal, a maximin in the Huber sense Neyman-Pearson detector based on the minimum distance between the signal and observations is designed. Explicit formulas for the power of detection and the false-alarm probability are derived. The maximin detectors are written out for the classes of nondegenerate, with a bounded variance and contaminated Gaussian noise distributions along with some numerical results on their performance.