A Zonotopic FDI Approach with LPV-based EKF in Autonomous Vehicles

This paper presents a fault detection and isolation approach for state estimation in autonomous vehicles, enhancing safety through the assurance of accurate measurement acquisition. Using a zonotopic method to guarantee robustness against unknown-but-bounded measurement noises, this study introduces a novel formulation of the Extended Kalman Filters, using the Linear Parameter-Varying modelling technique. Its performance is assessed alongside other state estimation methods. Experiments on a Renault Zoe (SAE level 3) enable a safety validation for the FDI process, finding the critical sensors for the state estimation algorithm. Copyright (c) 2024 The Authors.