Variability in decision-making and critical cue use by different road users at rail level crossings

Abstract

Collisions at rail level crossings (RLXs) are typically high-severity and high-cost, often involving serious injuries, fatalities and major disruptions to the transport network. Most research examining behaviour at RLXs has focused exclusively on drivers and consequently there is little knowledge on how other road users make decisions at RLXs. We collected drivers', motorcyclists', bicyclists' and pedestrians' self-reported daily experiences at RLXs for two weeks, focusing on behaviour, decision-making and information use in the presence of a train and/or activated RLX signals. Both information use and behaviour differed between road users. Visual information (e.g., flashing lights) was more influential for motorists, whereas pedestrians and cyclists relied more on auditory information (e.g., bells). Pedestrians were also more likely to violate active RLX warnings and/or cross before an approaching train. These results emphasise the importance of adopting holistic RLX design approaches that support cognition and behaviour across for all road users.Keywords: rail level crossings; grade crossings; decision-making; situation awareness Practitioner summary: This study explores how information use and decisionmaking at rail level crossings (RLXs) differs between road user groups, using a two-week self-report study. Most users make safe decisions, but pedestrians are most likely to violate RLX warnings. Information use (visual vs. auditory) also differs substantially between road user groups. IntroductionRail level crossings (RLXs; also called highway-rail grade crossings) are complex sociotechnical systems in which human (e.g., vehicle operators, pedestrians) and nonhuman (e.g., boom gates, flashing lights, trains, road vehicles) agents interact. RLX systems encompass diverse agents including train drivers, rail infrastructure, signals, signal operators, vehicle drivers, pedestrians, cyclists and motorcyclists (Read, Salmon, Investigation revealed the truck driver failed to detect the train, despite several passive and active warnings (for analysis, see Salmon, Read, et al. 2013). This has prompted research examining the functioning of RLX systems and warnings, to understand why some warnings are ineffective and how they might be improved. Reducing collisions requires understanding factors involved in not onlycollisions, but also user behaviour more broadly. Human factors are a major contributor to RLX collisions (Caird et al. 2002; Edquist et al. 2009), yet our understanding of influences on behaviour at RLXs remains poor. Previous research has relied on methods that provide limited insight into cognitive processes; for example, roadside observations (Neisser 1976; Plant and Stanton 2013a). Schemata are formed through experience and direct subsequent exploration and interpretation of our environment, meaning that information-seeking (and, consequently, decision-making based on that information) will be jointly determined by the RLX environment and our schema. Neisser (1976) argued that schemata have greatest inf...

Optimising the future of technology in organisations: A human factors perspective

“…Furthermore, as a passive level crossing, there were no boom gates that would otherwise have prevented the truck from entering the rail corridor. A series of design solutions that emerged as a consequence of the analysis have been subjected to empirical evaluation to identify the optimal solution in those circumstances where drivers become de-sensitised to passive level crossings (Beanland et al, 2015).…”

To stop or not to stop: Contrasting compliant and non-compliant driver behaviour at rural rail level crossings

“…CDM has been used to examine naturalistic decision making across a range of domains (Hoffman et al, 1998;Plant and Stanton, 2013;Stanton et al, 2007), including RLXs (Mulvihill et al, 2016). Although CDM was developed to study expert decisionmaking, it has been successfully adapted to examine decision-making while driving (Beanland et al, 2016;Stanton et al, 2007;Walker et al, 2009;Young et al, 2015). At the beginning of each interview, the participant was asked to describe a specific RLX encounter in as much detail as possible.…”

“…One option for future research would be to conduct a long-term self-report study of road users who commonly encounter passive RLXs; for instance, asking drivers to keep a diary record of how many RLXs they encounter each day and the events that occur. We previously used a similar method to explore how road users make stop-go decisions at RLXs, with the diary questions structured around CDM prompts (Beanland et al, 2016;Mulvihill et al, 2016). However, participants only completed the CDM if they saw a train on approach to the RLX, so the data yielded very few encounters with passive crossings.…”