Spectral analysis of stochastically sampled dynamic systems (Q5936219)

After describing some practical phenomena, the authors consider a dynamic stochastic model called an extended stationary Langevin model, which in fact is a Gaussian and stationary process \(x\) in \(\mathbb{R}^n\) (this is the signal), while the observed process \(y\) is an \(\mathbb{R}^n\)-valued linear function of the signal. The goal is to derive some conclusions about the spectral density of the signal process based on discrete observations. Both cases, when the observation times are deterministic and when they are random, are treated. The main attention is paid to discrete, that is, sampled models. A numerical illustrative example is given.