Estimation of the stepwise random disturbance parameters with the unknown moment of appearance

Abstract

We have carried out the synthesis of the joint algorithms for the estimation of the parameters of Gaussian random disturbance by the maximum likelihood method. Application of the generalization of local Markov approximation method and small parameter method makes it possible for us to find the asymptotically exact expressions for the characteristics of the specified estimates. We also establish the error range of the obtained theoretical formulas experimentally, by statistical computer simulation.