Vehicle–soil interaction: Testing, modeling, calibration and validation

•Stochastic soil–tire interaction models were calibrated and validated.•Local and global interval-based metrics were used in validation.•Physical and statistical models performed well using new time-domain test data. Although many studies have been conducted on different aspects of tire–soil interaction, little work has been done focusing on the uncertainties involved such as those of the mechanical properties of soil and the interfacial properties between the tire and soil. Even less, if any, work has been done on the validation of stochastic tire–soil interaction models using rigorous statistical methods. In this paper, a statistical framework, along with new vehicle–soil interaction test data, is used to build a stochastic metamodel from a simple physically-based tire–soil interaction model, to calibrate model parameters with uncertainties, to predict model responses with uncertainties, and to validate the models using four validation metrics: one local metric that measures the differences between test and model at each instant of time, and three global metrics that measure these differences but over the entire time period of vehicle motion. Results in using the metrics indicate that the models performed well.