Prognostic and Health Management Fundamentals: Particle-Filtering-Based Prognosis

Abstract:
Particle filters (PF) have been established as the de facto state-of-the-art in failure prognosis,
and particularly in the representation of the uncertainty that is associated to long-term
predictions. This talk explores some of the most important aspects associated to the problem
of failure prognosis, with special emphasis on the utilization of Particle Filter (PF) algorithms
for uncertainty representation, and outer feedback correction loops for uncertainty
management in prediction. Application examples will include problems associated to battery
management systems.