KALMAN FILTERING AND PARAMETER ESTIMATION IN THE ELEMENTARY GAUSSIAN PROCESS CASE

摘要: Abstract The parameter identification problem is discussed in the presence of additive coloured noise for stochastic processes with continuous time. Such are, example, stationary rational spectral density function. main tool to get Radon-Nikodym derivatives, distribution sufficient statistics exact and explicity solution Riccati equation Kalman filtering. Asymptotic results distributions are presented. method Novikov obtaining generalized, Ornstein-Uhlenbeck process detail. observed has following form: where both θ(t), e(t) type unknown parameters α, β, which have be estimated.