Modelling driver behaviour towards innovative warning devices at railway level crossings

Abstract

Improving safety at railway level crossings is costly and as funds are often limited, it is important to search for cost-effective, evidence-based solutions. The effect that the many existing alternative systems have on driver behaviour is not always known. This paper compares driver behaviour towards two novel warning devices (rumble strips and in-vehicle audio warning) at railway level crossings with two conventional warning devices (flashing light and stop sign). Regression models were developed to reflect driver's responses towards the four different types of devices based on data collected from a driving simulation experiment. The regression models include a binary choice model for predicting the probability of a driver stopping or driving through a railway crossing, as well as mixed regression models for predicting the moment at which a driver will produce specific behavioural responses before stopping at a crossing (e.g. initiation of accelerator release and application of foot-pedal brake). Violation results indicated the active systems produced much higher levels of driver compliance than passive devices. Contributing factors, such as age, gender, speed and types of warning devices were found significant at different approach stages to the level crossings. With the application of such behavioural models and traffic conflict techniques in microscopic simulation tools, traffic safety indicators, such as collision likelihood and time-to-collision can be estimated. From these, relative safety comparisons for the different traffic devices are derived.