State observers applied to vehicle lateral dynamics estimation: a comparison between Extended Kalman filter and Particle filter
Résumé
The vehicle dynamics motion is essential to the vehicle stability and safety character. Some active safety systems have been integrated as standard options in our quotidian cars. These safety systems use simple vehicle parameters which are available on the vehicle on-boarded informatics system. However, the complex safety system or the advanced control unit onboarded on the car may need more precise parameters describing the vehicle dynamics in an accurate way. The tire/road force and sideslip angle, that are not available on the onboarded informatics system of standard cars and can not be measured by inexpensive sensors, can give direct description of the vehicle dynamics state. This study introduces a method to estimate these dynamics parameters which are valuable to the development of more complex and powerful drive assistance system. In this article, we present a comparison between two nonlinear state observer methods, one based on Extended Kalman filter and another based on Particle filter technique. We compare the robustness of these two observers through experimental validation performed with our laboratory experimental vehicle.