Recognition of statistically defined signals along with unknown signals

Abstract

Some practical peculiarities of methods to recognize random signal are considered for non-traditional cases when there are both signals with unknown probabilistic characteristics and statistically defined signals are given for recognition. The novelty of the present work is that it considers the signal recognition problem solving with employing methods based on different probabilistic models like auto regression processes, mixtures of standard distributions and expansions of random signals with respect to orthogonal functions. Consideration is given to results of the applied signal recognition problem solving in the field of radar and radio monitoring.