Estimation of the Parameters and Recognition of Stochastic Processes

Abstract

While processing the information there arises the necessity to solve such important problems as — recognition, detection, estimation of the parameters, determination of the momenta of time changes in the properties of stochastic signals. When solving these problems the correlation methods for Gaussian models have been considered. A new approach based on Markov models of stochastic processes has been developed [1]. On the basis of the above model the solution of such problems as filtering, extrapolation, interpolation has been obtained [1], [2]. In the present communication the problem of estimation of the parameters and that of recognition of multidimensional stochastic processes under noise conditions is being considered on the basis of the theory of conditional Markov processes [3].