An Algorithm for Parameter Identification of Semi-Empirical Tire Model

Vehicle tire performance is an important consideration for vehicle handling, stability, mobility, and ride comfort, as well as durability. All forces exerted on the vehicle, except aerodynamic forces, are transferred to the contact regions between the road and tires. As one of the most famous semi-empirical tire models, the Magic Formula (MF) model is widely used in vehicle ride comfort and handling stability simulations because of its ability to characterize the dynamic characteristics of tires. However, it is difficult to quickly and accurately identify the MF model that contains many parameters and highly nonlinear characteristics. This article introduces a homotopy optimization methodology to identify the MF tire model parameters based on weighted orthogonal residuals, with a morphing parameter used to lead the algorithm to the optimal global solution and avoids local convergence. The idea of weighted orthogonal distance regression (WODR) was used instead of ordinary least squares (OLS) to accurately identify the parameters and make the identification errors smaller and more evenly distributed. The results show that the homotopy optimization based on WODR has good precision and high efficiency.