Extended Kalman Filter Based Lateral Velocity and Yaw Rate Estimation for a Vehicle with Nonlinear Tire Model

The extended Kalman filter (EKF) is adapted to estimate the state of a nonlinear vehicle single-track model. EKF’s basic idea relies on an iterative linearization of the dynamic equations around the current state estimate and nominal control. For the problem of assisted vehicle lateral dynamic control, lateral velocity and yaw speed are important to make a rough guess since they are unavailable using standard sensors onboard. The accurate state estimate, guaranteed by the EKF in the presence of noisy measurements, can then be easily used in any control state feed-back framework. The nonlinear model used for state estimation in this paper does not assume linear model for tire friction forces, instead it is based on the highly nonlinear reliable Pacejka model. The vehicle tires’ slip angles required by the Pacejka model are derived using very few simplifying hypotheses compared to other papers in the literature. Numerical simulation, considering different challenging scenarios, is conducted to assess the EKF accuracy and robustness against parameters uncertainties. The results are illustrated to demonstrate the filter’s effectiveness.