Warranting traffic signals on the basis of proximity of railroad grade crossings

There were two concurrent paths in this research. Path 1 estimated the probability, p_{vehicle/train}, that a vehicle will be in the train dynamic envelope given that a train arrives. This was done by an extensive simulation effort that included some data collection on relevant gap acceptance patterns. Path 2 identified the exposure factors, selected a risk metric, and established a target or threshold for the risk, which is expressed as an upper bound on p_{vehicle/train}. As a result of the investigation, the warrant is based on the principle that the expected fatality rate at the crossing should not exceed rates found at unsignalized intersections. The two paths converged by expressing the constraint p _{vehicle/train} ≤_(where_is in terms of traffic terms in common use - e.g., volumes, composition, track location distances). This constraint led to the warrant, which was then validated with data acquired or collected for the purpose. The draft warrant was prepared as a result of the undertaking, was endorsed by the National Committee on Uniform Traffic Control Devices, and is currently in the formal review process, culminating in final edits and a decision on inclusion in the next edition of the Manual on Uniform Traffic Control Devices.