Estimation of Vehicle Dynamics in Driving Simulator

초록

A driving simulator is a cost-effective research and development tool in ultimately safe environment with less effort so that it is ideally suited for designing self-driving cars. This paper introduces a low-cost driving simulator that was built by our previous researches based on a commercial game GTA V. A method for the system identification of vehicle dynamics is suggested for the efficient use of simulator technology prior to the field test of a vehicle on realistic roads. A simple form of bicycle model is employed in the study for the system identification of transfer function with the vehicle dynamics parameters. The system identification is performed using the Least Squares Estimation (LSE) technique to estimate the transfer function of vehicle dynamics. The required data were generated and acquired through the GTA V driving simulator, and then applied to the developed vehicle dynamics estimator for testing.