System State Estimation by Particle Filtering for Fault Diagnosis and Prognosis
Abstract
Efficient diagnosis and prognosis of system faults depend on the ability to estimate the system state. In many real applications, the system dynamics is typically characterized by transitions among discrete modes of operation, each one giving rise to a specific continuous dynamics of evolution. The estimation of the state of these hybrid dynamic systems is a particularly challenging task because it requires tracking the system dynamics corresponding to the different modes of operation. In the present paper a Monte Carlo-based estimation method, called particle filtering, is illustrated with reference to a case study of a hybrid system from the literature.