Modal estimation in hybrid systems

Abstract

In hybrid systems, estimation is made difficult by the nonlinear equations of state evolution. Several algorithms for generating an approximation to the conditional-mean estimator have been proposed. These are satisfactory in many cases, but because of the way the observations are fused, the estimate of the modal component of the state is often less good. This paper proposes a novel algorithm for state estimation and compares it with some in current use. © 2000 Academic Press.