Effect of Pavement Marking Retroreflectivity on the Performance of Vision-Based Lane Departure Warning Systems

Abstract

Vision-based lane departure warning systems (LDWS) use images collected by small cameras to track the pavement markings and warn drivers when their vehicles perform unintended drifts out of the travel lanes. Pavement marking visibility at night is a function of the marking's retroreflectivity. During rain, pavement markings are covered by water, reducing the retroreflectivity and adversely affecting the visibility of the markings at night. This study examines the effects of the marking's retroreflectivity on the performance of LDWS, based on tests conducted in the field and in a simulated rain facility. The results indicate that improving the retroreflectivity attributes of commonly used pavement marking materials such as thermoplastic and paints improves the performance of LDWS for rain conditions at night. In addition, this performance improves significantly with the use of enhanced types of pavement markings that have high visibility during rain.