A Prognostic Approach Based on Particle Filtering and Optimized Tuning Kernel Smoothing

Abstract

This paper proposes a novel approach based on a Particle Filtering technique and an Optimized Tuning Kernel Smoothing method for the prediction on the Remaining Useful Life (RUL) of a degrading component. We consider a case in which a model describing the degradation process is available, but the exact values of the model parameters are unknown and observations of historical degradation trajectories in similar components are unavailable. A numerical application concerning the prediction of the RUL of degrading Lithium-ion batteries is considered. The obtained results show that the proposed method can provide a satisfactory RUL prediction as well as the parameters estimation.