Optimal state estimation in high noise

Abstract

The problem is examined of estimating the state of a linear dynamical system in the presence of high measurement noise. It is concluded that optimal filter design may be simplified to the extent that it need not depend on the solution of a matrix Riccati differential equation, but only on the solution of a matrix linear differential equation. For a related problem, that of estimating a signal s(t) given noisy measurements s(t) + w(t) where the noise is large and the covariance of s(t) is known, optimal filter design is immediate. Show more