A train speed measurement and arrival time prediction system for highway-rail grade crossings

A key component to the safe and efficient operation of traffic signals near highway-rail grade crossings (HRGCs) is an accurate estimate of a given train's arrival time at each crossing. Based on this information, current signal preemption strategies for traffic signals near the HRGCs could be improved with respect to driver safety, pedestrian safety and network delay. This paper introduces a system for train speed measurements and arrival time prediction at HRGC using data from Autoscope video detection and a Doppler radar detector. The two detectors were used in parallel because each, on its own, has a number of issues that preclude accurate detection of simultaneous trains (e.g., trains on different tracks in the detection zone at the same time). A multiple-track test bed system in Lincoln, Nebraska was chosen for train speed data collection. A number of potential kinematic equation-based and statistical models were calibrated to predict train arrival time at the HRGC. These were then analyzed and the best models, based on absolute errors, were identified. The regression models proved to be more accurate than the kinematic models.