Continuous-time stochastic model simplification

Approximation of high-order and time-varying linear continuous-time Gaussian models by low-order and time-invariant models is considered. The recent work of Baram and Be'eri pertaining to the discrete-time systems is extended to continuous-time system models. The model simplification is carried out by maximizing the probabilistic ambiguity between the actual system and the approximate model. Differences between the continuous-time and the discrete-time model simplification problems are examined. The case when the observation noise covariance of the approximate model differs from that of the actual system presents technical difficulties not present in the discrete-time version.