PROMOTING SAFETY AT RAILROAD CROSSINGS BY REDUCING TRAFFIC DELAYS

This study develops an artificial neural network traffic control algorithm in order to optimize traffic delays around highway railroad crossings. The 2-step algorithm first designs a proper preemption phase plan, and then finds the optimized phase length. The aim of the preemption plan design is to maximize safety at grade crossings. This can be achieved by designing the preemption plan so that highway traffic will be prevented from queuing on the grade crossing intersection. The optimized process will use as objective function traffic delays at intersections surrounding the grade crossing area. That function will be approximated and represented by neural network. Next, mathematical algorithms are employed to get the optimized length of phases so that total delays can be minimized. This research uses the CORSIM simulated traffic network package to conduct analysis and determine results.