Digital signals processing using non-linear orthogonal transformation in frequency domain

The rapid progress of computer technology in recent decades led to a wide introduction of methods of digital information processing practically in all fields of scientific research. In this case, among various applications of computing one of the most important places is occupied by digital processing systems signals (DSP) that are used in data processing remote solution tasks of navigation of aerospace and marine objects, communications, radiophysics, digital optics and in a number of other applications. Digital Signal Processing (DSP) is a dynamically developing an area that covers both technical and software tools. Related areas for digital signal processing are theory information, in particular, the theory of optimal signal reception and theory pattern recognition. In the first case, the main problem is signal extraction against a background of noise and interference of a different physical nature, and in the second - automatic recognition, i.e. classification and signal identification. In the digital processing of signals under a signal, we mean its mathematical description, i.e. a certain real function, containing information on the state or behavior of a physical system under an event that can be defined on a continuous or discrete space of time variation or spatial coordinates. In the broad sense, DSP systems mean a complex algorithmic, hardware and software. As a rule, systems contain specialized technical means of preliminary (or primary) signal processing and special technical means for secondary processing of signals. Means of pretreatment are designed to process the original signals observed in general case against a background of random noise and interference of a different physical nature and represented in the form of discrete digital samples, for the purpose of detecting and selection (selection) of the useful signal and evaluation characteristics of the detected signal. A new method of digital signal processing in the frequency domain is proposed, using some properties of nonlinear orthogonal transformations. The theoretical bases of this method and the results of the statistical modeling, proving its efficacy in suppressing concentrated interference in the communication channels.