Joint Tire-Stiffness and Vehicle-Inertial Parameter Estimation for Improved Predictive Control

• Published in: 2021 American Control Conference (ACC) pp. 186 - 191
• Main Authors: Berntorp, Karl, Quirynen, Rien, Vaskov, Sean
• Format: Conference Proceeding
• Language: English
• Published: American Automatic Control Council 25-05-2021
• Subjects: Computational modeling; Estimation; Parameter estimation; Production; Simulation; Tires; Uncertainty
• ISSN: 2378-5861
• DOI: 10.23919/ACC50511.2021.9482635

This paper presents a method for online estimation of linear friction (i.e., tire stiffness) and inertial parameters (i.e., mass and inertia) using sensors readily available from the CAN bus in production vehicles. We treat the tire stiffness as a time-varying Gaussian disturbance acting on the vehicle, and the inertial parameters are modeled as nearly constant parameters with large initial uncertainty. We leverage particle filtering and the marginalization concept to estimate in a computationally efficient way the tire-stiffness and inertial parameters, together with the vehicle state. We integrate the estimator with a nonlinear model-predictive controller (NMPC) and evaluate the efficacy of the estimator in closed-loop control.