Robust Fault Detection for Vehicle Lateral Dynamics: A Zonotope-based Set-membership Approach

Abstract

In this work, a model-based fault detection layout for vehicle lateral dynamics system is presented. The major focus in this study is on the handling of model uncertainties and unknown inputs. In fact, the vehicle lateral model is affected by several parameter variations such as longitudinal velocity, cornering stiffnesses coefficients and unknown inputs like wind gust disturbances. Cornering stiffness parameters variation is considered to be unknown but bounded with known compact set. Their effect is addressed by generating intervals for the residuals based on the zonotope representation of all possible values. The developed fault detection procedure has been tested using real driving data acquired from a prototype vehicle.

Vehicle Lateral Velocity and Lateral Tire-road Forces Estimation Based on Switched Interval Observers

“…The provided compact set is consistent with the measurement, the vehicle model and the bounded uncertainties. Recently, interval estimation based-approach has attracted the interest of many researchers and important results are available in the literature for different class of dynamical systems [3], [4], [6], [11], [14]. They were originally developed in [5] for the estimation of biological systems subject to unknown uncertainties.…”