1. Introduction
Traditional automatic radar processing systems while working quite well in simple situations are ineffective in many target environment or if nonstationary and complex noise and jamming affect the radar. Unfortunately, methods for sufficiently effective automatic radar processing does not exist. There is an advanced and comprehensive theory for radar information processing [1]. However, it is based on many assumptions that obviously are not met in practice. Examples of such assumptions are stationarity and normal distribution of clutter, separate target handling, etc.