01664157

Component

This paper uses micro-level vehicle profile data extracted from radar-based sensors to: (1) identify the key variables associated with gate violations at highway-rail grade crossings (HRGCs), (2) develop prediction models for gate violations at HRGCs, and (3) examine the relationship between model accuracy and the key input variables. A data set of 256 vehicle-train events was collected at an HRGC test bed in Lincoln, Nebraska. Among them, 76 events are gate violations and 180 events are non-violations. The vehicle profiles, over a 300 foot distance from the HRGC stop line, were discretized at 10 foot increments. Each increment had 3 variables associated with it. Two tree-based ensemble techniques, the bootstrap forest, and the boosted tree, were applied to determine the relationship among the input variables and the occurrence of gate violation. It was found that once a vehicle is within 190 feet of the HRGC stop line, the model was approximately 80 percent accurate in predicting a gate violation. Influential factors that were identified include speed, time-to-gate, and acceleration over a distance to the stop line. It was found that as the vehicles came closer to the HRGC, and more information was contained in the vehicle profile, the prediction error decreased. It should be noted that with the advent of vehicle profile data collection, the tree-based ensemble techniques are ideal candidates for safety studies because they can utilize the highly non-linear vehicle profiles and relate these to safety surrogate metrics.

This paper was sponsored by TRB committee AHB60 Standing Committee on Highway/Rail Grade Crossings. Alternative Title: Predicting Highway–Rail Grade Crossing (HRGC) Gate Violations Using Tree-Based Ensemble Techniques

18-05343

English

4p

2018

Transportation Research Board 97th Annual Meeting

Digital/other

Figures; References (12)

Gates; Grade crossing protection systems; Highway safety; Mathematical prediction; Moving violations; Railroad grade crossings; Vehicles

Lincoln (Nebraska)

Highways; Railroads; Safety and Human Factors

Transportation Research Board Annual Meeting 2018 Paper #18-05343

TRIS, TRB, ATRI

Jan 8 2018 11:21AM