Abstract
Assuming a five degree of freedom non-linear bicycle model, the transient response of the vehicle is simulated assuming the presence of stochastic stationary and time varying noise and disturbance properties. Various combinations of linear and non-linear tyre model in the plant and the Extended Kalman Filter (EKF) design are simulated in order to investigate the effects of plant uncertainty on EKF estimation performance. It is found that the EKF is only able to generate state estimates that track the true response when the same tyre model is used in both the plant and EKF design. Therefore, varying approaches from the literature for robust EKF design using parameter identification are discussed with a view to future work.
Keywords: Extended Kalman Filter Dynamic Systems