Data-driven linear parameter-varying modelling of the steering dynamics of an autonomous car

Developing automatic driving solutions and driver support systems requires accurate vehicle specific models to describe predict the associated motion dynamics of vehicle. Despite mature understanding ideal dynamics, which are inherently nonlinear, modern cars equipped with a wide array digital mechatronic components that difficult model. Furthermore, due manufacturing, each car has its personal characteristics change over time. Hence, it is important develop data-driven modelling methods capable capture from data all relevant aspects in model directly utilisable for control. In this paper, we show how Linear Parameter-Varying (LPV) system identification can be applied reliably personalised steering an autonomous based on measured data. Compared other nonlinear techniques, obtained LPV powerful controller synthesis framework.