Driver’s Over-Trust on Advanced Driver Assistance Systems on Passively Protected Railway Crossings

Larue, Gregoire S.; Rakotonirainy, Andry · 2022 · Crossref

DOI: 10.54941/ahfe100747

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study addresses the safety risks associated with passively protected railway crossings in Australia, where upgrading infrastructure to active protection is often economically unfeasible. While Advanced Driver Assistance Systems (ADAS) offer a cost-effective solution to improve safety, the authors investigate the potential for driver over-trust. Specifically, the research examines whether drivers inappropriately allocate trust to non-fail-safe ADAS, treating them as equivalent to fail-safe active infrastructure, thereby neglecting necessary safety precautions. To test this hypothesis, the researchers conducted a driving simulator study with 58 participants. Participants were exposed to three types of ADAS interventions—visual in-vehicle warnings, audio in-vehicle warnings, and on-road flashing beacons—alongside baseline conditions involving passive and active crossings. The experimental design utilized a within-subject analysis where participants drove scenarios with and without ADAS. Data collection focused on stopping compliance, gaze patterns toward rail tracks, and approach speed 20 meters from the crossing. Statistical analysis employed Generalised Linear Mixed Modelling to account for repeated measures. The results indicated that while ADAS improved stopping compliance when a train was approaching (increasing to 96%, comparable to active crossings), it significantly reduced compliance when no train was present. Stopping compliance dropped from 79% in the baseline passive condition to 61% with ADAS. Furthermore, gaze compliance decreased from 97% to 93%, indicating drivers checked the tracks less frequently. Approach speeds also increased when the ADAS indicated no train was present, with non-compliant participants reaching speeds similar to those observed at active crossings without trains. These findings demonstrate that a significant proportion of drivers treated the ADAS as the primary control rather than a complementary aid, failing to stop or visually verify the crossing status when the system provided no warning. The study concludes that current ADAS designs for passive crossings risk inducing dangerous over-trust because they lack fail-safe modes and do not clearly communicate system status or limitations. Drivers mistakenly equate the reliability of ADAS with that of traditional active signage, leading to riskier behaviors such as higher speeds and reduced vigilance. The authors argue that the assumption that drivers will continue to rely on passive signage as the primary control is invalid. Consequently, they recommend radically different Human-Machine Interface designs for ADAS to prevent drivers from viewing the technology as equivalent to active protection, ensuring it remains a supplementary tool rather than a primary safety mechanism.

Key finding

A significant proportion of drivers over-trust ADAS at passive railway crossings, resulting in reduced stopping compliance, higher approach speeds, and fewer visual checks when no train is present.

Methodology

simulator

Sample size: 58

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• rail grade crossings
• trust calibration