Abstract
The benefit of previewing the “future” road inputs in improving the performance of active suspensions of a vehicle has been demonstrated in many publications. It is shown that smoother and more effective control is possible when the preview length is increased up to a certain limit. Most of these studies assume perfect preview measurements in their numerical simulations. In a real scenario however, the presence of uncertainties can not be neglected. The uncertainties can be due to sensor measurement noise or false objects on the surface of the road. In order to gain an insight to the problem this paper addresses the influence of preview uncertainty on the performance of a preview controller. The stochastic optimal control scheme for preview control of a quarter-car suspension model is briefly presented. Then results of numerical simulations, which are carried out for various levels of uncertainty in the preview and for different preview lengths, are compared. These comparisons show that noise/uncertainty can be an issue with an adverse effect that sometimes increases with preview length. In other words, it is shown that controllers with longer preview can be more susceptible to noise. However the limits of tolerable noise are not very restrictive and the well-known benefits of preview can still be achieved if the measurement errors are maintained below a limited level.