Signal Detection Algorithms Based on Non-Parametric Estimates of Density Function

This paper presents a novel approach to design radar signal detection algorithms that are applicable when a priori information is limited. The problem is formulated as testing hypothesis on the kind of density function. A novel method that allows to adopt permutation test in a practical algorithm is suggested and researched. The developed new adaptive algorithm is based on non-parametric kernel estimates of the density function. The results are useful for applications of signal detection in surveillance and remote sensing radar systems.