On the vehicle dynamics prediction via model-based observation

Accurate knowledge of the vehicle dynamics response is a critical aspect to improve handling performance while ensuring safe driving at the same time. However, it poses a challenge since not all the quantities of interest can be directly measured due to cost and/or technological reasons. Therefore, several methods have been developed relying on physical models that map the relationship between these uncertain quantities and other variables that are directly measurable via the onboard sensors. This approach is referred to as model-based estimation, and it is usually solved via Kalman Filtering (KF). The accuracy that can be achieved is tightly connected with the model and the estimation algorithm selected by the designer. In this paper, models with varying levels of fidelity and different KF-based estimators are compared in order to shed some light on the appropriate construction of a model-based observer among the large body of research present in the literature. Recent nonlinear estimation algorithms including the Unscented Kalman Filter (UKF) and the Cubature Kalman Filter (CKF) are contrasted with each other and against the standard Extended Kalman Filter (EKF) on experimental data available from a public data set that uses an instrumented Ferrari 250 LM Berlinetta GT as a test bed.

Automated summary

Accurate knowledge of vehicle dynamics response is crucial for improving handling performance and ensuring safe driving. However, due to cost and technological limitations, not all quantities of interest can be directly measured. Model-based estimation methods, such as Kalman Filtering (KF), have been developed to map the relationship between uncertain quantities and measurable variables. This paper compares models of varying fidelity and KF-based estimators to guide the construction of a model-based observer. Nonlinear estimation algorithms, including the Unscented Kalman Filter (UKF) and Cubature Kalman Filter (CKF), are contrasted with the standard Extended Kalman Filter (EKF) using experimental data from a public dataset.