State, parameter and input observers based on multibody models and Kalman filters for vehicle dynamics. (16th June 2021)

• Record Type: Journal Article
• Authors: Rodríguez, Antonio J.; Sanjurjo, Emilio; Pastorino, Roland; Naya, Miguel Ángel
• Is Part Of: Mechanical systems and signal processing. Volume 155(2021)
• Publication Date: 2021-06-16
• Subjects: Multibody dynamics -- Kalman filter -- State observer -- Force estimation -- Parameter estimation -- Digital twin -- Virtual sensing -- Vehicle dynamics -- tire-road friction coefficient; Structural dynamics -- Periodicals; Vibration -- Periodicals; Constructions -- Dynamique -- Périodiques; Vibration -- Périodiques; Structural dynamics; Vibration; Periodicals; 621
• DOI: 10.1016/j.ymssp.2020.107544
• Languages: English
• ISSNs: 0888-3270

Abstract

Abstract: The aim of this work is to present a novel accurate estimator for vehicle dynamics. Following the multibody dynamics approach, a vehicle can be modeled with a high level of detail including non-linear dynamics. As a consequence, a rich simulation data-set is available for engineering analysis, richer than with vehicle analytical models. The proposed novel estimator is a new form of a dual Kalman filter. The first filter uses an indirect extended Kalman filter (i.e. the errorEKF) incorporating force estimation and using a vehicle multibody model. The second filter is an unscented Kalman filter (UKF) used to increase the accuracy of the errorEKF by estimating uncertain modeling parameters such as the mass of the vehicle and the tire-road friction coefficient. The performance of the proposed state-parameter-input (SPI) observer is tested in a simulation environment. The performance of the observer is demonstrated using two maneuvers, out of which one covers aggressive driving conditions. The results show that the new observer estimates with high accuracy the variables of interest for vehicle dynamics, such as the tire forces.