Tykhonov V. Development of Linear Prediction Models for Non-Gaussian Stochastic Processes

Object - non-Gaussian stochastic processes and signals and their statistical processing. The purpose - design of generic models of linear prediction of statistically connected non-Gaussian processes and analysis of possibilities for their application for improving efficiency of various radio-electronic systems. Methods - theory of linear systems, linear differential equations, linear operators, statistical simulation, applied analysis of stochastic processes. Results - new generic models of linear prediction, parametrical estimations of higher-orders spectrums, synthesis theory of generic Wiener filters, lattice structures, compensators for inter-symbol interference, recognition methods for non-Gaussian signals. It is introduced - in research and development projects, in design of data transmition devices, in education. Sphere of application - analysis and processing of non-Gaussian processes