In previous work a linear model of driver steering control was developed which takes account of human sensory dynamics and limitations. In this paper various approaches to modelling a driver’s control of a nonlinear vehicle are compared. In contrast to research focussed on modelling the optimal driver, the aim of this work is to develop a realistic model of driver steering behaviour. Simulations were run to compare various nonlinear state estimators and controllers. In general a trade-off was found between simulation time, which could also represent mental load, and controller performance. Experiments are planned to compare the results of these simulations against measured steering behaviour from human drivers.