Vehicle Dynamics Parameter Estimation Methodology for Virtual Automated Driving Testing

The development and validation of automated driving vehicles requires an intensive testing process that can only be performed efficiently with the support of virtual simulators. In the simulation process, the vehicle dynamics behaviour influences the testing of controllers and decision algorithms. Although the use of vehicle dynamics models is well described in the literature, parameter estimation is usually a costly process. In order to achieve a straightforward estimation of the parameters, this work proposes a combination of static and dynamic measurements and the implementation of optimization methods to estimate the parameters. Initially, the vehicle mass and position are calculated from the vehicle weight and the tyre parameters are measured on a dedicated test bench. The transmission, torque converter and other tyre parameters are calculated from data collected during dynamic maneuvers. Finally, a differential evolution method is applied to minimise the errors between the model and the data collected from the real vehicle when subjected to the same control inputs. This work makes an important contribution to accelerate the virtualisation of vehicle dynamics, providing the community with a comprehensive open-source vehicle dynamics model. Furthermore, the proposed methodology has proven to be an efficient way to estimate vehicle dynamics parameters and can be used to test specific maneuvers in the simulated environment.