It is imperative to have accurate tire models when trying to control the trajectory of a vehicle. With the emergence of autonomous vehicles, it is more important than ever before to have models that predict how the vehicle will operate in any situation. Many different types of tire models have been developed and validated, including physics-based models such as brush models, black box models, finite element-based models, and empirical models driven by data such as the Magic Formula model. The latter is widely acknowledged to be one of the most accurate tire models available; however, collecting data for this model is not an easy task. Collecting data is often accomplished through rigorous testing in a dedicated facility. This is a long and expensive procedure which generally destroys many tires before a comprehensive data set is acquired. Using a Vehicle Measurement System (VMS), tires can be modeled through on-road data alone. This reduces the time and cost significantly and does not require destroying multiple tires. Previous works regarding this parameter identification method have used only the basic versions of the Magic Formula model-pure longitudinal slip and lateral sideslip-but the Magic Formula model also includes combined slip conditions as well. To accurately mimic the tire forces, especially in safety critical situations for autonomous vehicles, combined slip tire models are necessary. The longitudinal slip, sideslip angle, tire forces, and tire moments are measured and calculated using a VMS during normal and extreme driving scenarios. The data is then used to identify the parameters for the 1989 Pacejka model for both pure slip and combined slip scenarios. These models are then implemented and validated with a full vehicle dynamic model.
