Optimal parameter estimation in semi-empirical tire models

Abstract

Semi-empirical tire models are mathematical models, the parameters of which are identified after a process of error reduction to fit experimental data obtained in the laboratory. In this process, the algorithms used for estimating the model parameters are usually based on nonlinear least-squares fitting methods, in which only vertical residuals between the model and the test points are considered. Although extensively utilized, this type of fitting implicitly considers that errors in the slip data (horizontal residuals) are either nonexistent or negligible, which is not true. This paper introduces a new methodology to the identification of semi-empirical tire model parameters based on weighed orthogonal residuals, which takes into account possible errors inherent in the test measurements of dependent and independent variables. The results show that the methodology offers a reliable parameter identification providing an even fitting for all the zones of the mathematical semi-empirical tire model.