Written by Grégoire Larue, Research Fellow at ACRI. An abridged version of this article was published in Track & Signal July-September 2016.
In this issue of Track + Signal, I would like to discuss research into the effects of extended waiting times on driver behaviour at level crossings. This research, funded by the Australasian Centre for Rail Innovation (ACRI), was commissioned to better understand the situation at congested level crossings, with a focus on both the rail and the road perspectives. The aim was to find out whether there is a waiting time beyond which drivers are more likely to undertake risky behaviour at level crossings, and then use this knowledge to inform the standardisation of waiting and warning times at congested level crossings.
Level crossings during peak hours
Before I present our study in detail, some context on level crossings and driver behaviour is needed.
A growing population has resulted in road congestion at level crossings on Australian metropolitan train networks worsening in recent years. This congestion issue is particularly acute during peak periods when both road traffic and the frequency of trains are at their maximum.
Increases in train traffic leads to an increase in the number of level crossing closures, however this is only one of the factors leading to road traffic disruption and congestion.
Urban level crossings are characterised by a large variety of train types, ranging from metro (express or not), regional and freight trains. These trains travel at different speeds, which result in a large variability in the time taken by trains to reach the crossing from the activation of the flashing lights at the crossing. Level crossings are activated at a distance that ensures the fastest trains do not arrive at the crossing before a minimum warning time, which is defined in the Australian standard (AS1742.7). This ‘warning time’ provided to drivers can be as long as two minutes. The amount of time level crossings are closed also increases because of the lack of time available to safely re-open the crossing between consecutive trains, leading to extended closures for multiple trains at peak times.
A combination of these factors can result in level crossing closures being longer than they may need to be.
Road user behaviour at congested level crossings
Drivers’ decision making is affected by the amount of time that the driver needs to wait at the crossing. It is well known that excessive waiting times can lead to frustration and non-compliant behaviour by motorists, including driving through flashing lights, driving around the boom gates, and stopping on the yellow crossing marking.
These driving behaviours can result in significant safety risks. Although collisions at level crossings are relatively infrequent, the risk is high due to the number of crossings in Australia and the magnitude of traffic flowing through these crossings, as well as the severity of collisions between trains and road vehicles. The current increase in congestion is likely to increase this risk.
While the activation of level crossings follows the standard (by ensuring warning times are above the minimum legal requirement) and ensures the safety of compliant road users, improvements in the timing of activation are necessary to reduce the likelihood of driver violations and improve road and rail safety.
With a view of standardising waiting time at level crossings, the following questions should be answered: What would be an effective threshold of time to ensure a minimum amount of violations from road users? Is it the overall waiting time, or the fact that drivers are stopped at level crossings for a long time before the train actually arrives contributing significantly to level crossing violations? Answering such questions will allow those who operate and maintain level crossings to establish the most appropriate warning time for the location, as well as determining the level of treatment necessary to improve safety at these level crossings.
Our driving simulator experiment
The correlation between waiting time and road rules violations is the question that ACRI decided to tackle in this project: how long are motorists prepared to wait at level crossings before undertaking risky behaviour?
We tested the effect of different waiting times on driver decisions’ and risk taking behaviours on an advanced driving simulator (pictured). Participant drivers (62 males) were driving a course that involved navigating through a level crossing under time pressure. They drove the course six times and experienced various waiting times at the crossing, from the minimum necessary to let a train traverse the crossing, to ten minutes with multiple consecutive closures of the crossing and a variety of road traffic congestion situations.
To investigate the effect of the time taken by the train to arrive at the crossing, participants were divided into three groups, with each group experiencing a different warning time. For the first group, the train was travelling at the line speed and the warning time was the minimum required by the standard for the simulated crossing (28s). The second group encountered the path of a slower train, resulting in a 50s warning time (which is the average value we observed at congested level crossings in the Brisbane area). For the last group, the train stopped at a nearby station while the crossing was activated and resulted in a 75s warning time, which is the highest frequent value we observed at the Brisbane sites.
Figure 1: View from the control room of the Centre for Accident Research and Road Safety – Queensland’ advanced driving simulator used in our study. We replicated an active level crossing with boom barriers with the road and rail traffic of the morning peak, as shown on the screens.
With the use of time pressure on participants, we managed to recreate violations in our simulated environment. The violations observed are similar to the ones we have observed at a congested level crossing in the Melbourne area:
- Traversing the crossing when the flashing lights are activated and the boom gates are not down;
- Entering the level crossing when the boom gates are raising but the flashing lights are still active;
- Stopping on the yellow box at the level crossing; and
- Reversing on the crossing to avoid being stuck on the crossing
Our experiment showed that increasing waiting times (the total time a driver is stationary for) resulted in increased driver frustration and increased likelihood to engage in risky behaviour at level crossings.
The effect of waiting time was more pronounced when drivers were stopped at the crossing for longer times. This demonstrates that it is not only the level crossing closure time that matters. Other factors such as road congestion and the need to stay at the crossing for multiple activations of the crossing were factors that had high effect on drivers’ frustration and engagement in risky behaviour.
In order to evaluate the potential effects of reducing waiting times at level crossings, some participants experienced decreasing waiting times at the crossing, rather than increasing waiting times. This revealed further effects of interest when considering potential countermeasures, and how quickly drivers can adapt their behaviour to improved crossing conditions.
Understanding the reasons for the behaviour of level crossing users is a vital element in developing interventions aimed at modifying behaviour to improve safety. To assist with understanding driver motivations we obtained participant feedback at the end of their simulated drive.
Participants reported a high level of agreement with the importance of abiding by crossing rules. This related to participants generally believing it was not acceptable to cross at level crossings when the boom gate is down and the lights are flashing, nor breaking crossing rules as long as the train is not in sight.
However, despite this, violation rates remained high during the driving task. Participants reported a variety of contributing factors in their decisions to violate level crossing rules during their driving task that are discussed further in the report.
The repetition of driving through the crossing – which was used to create familiarity with the crossing and its congestion – proved very effective in leading to impatience/frustration while waiting, particularly when participants had to wait at the crossing for multiple activations.
Conclusions
This research provides insights into the effects of congestion at level crossings as well as an evaluation of the amount of time motorists are ready to spend waiting at crossings before violating road rules and entering into risky driving behaviour.
The outcomes of this project have shown the relevance of configuring signalling systems to operate at standard warning times for crossings before extended waiting times are experienced by motorists, at a threshold below that which generates significant risky behaviour. For level crossings which are already leading to risky behaviours, the effects of any improvements in the length of waiting times are discussed to provide an indication of how quickly operators might expect an improvement in driving behaviour to be observed, following a reduction in waiting time.
The full report is available on the ACRI Participant portal for current Participants of ACRI.
References (for more information on this research)
Larue, G.S, Blackman, R., Freeman, J., & Rakotonirainy, A. (2016). Impact of Waiting Times on Risk and Standardisation of Waiting Times. Australasian Centre for Rail Innovation report LC7-8.
Miska, M., Qian, G. Larue, G.S., & Rakotonirainy, A. (2015). Detailed Study of Current Rail Level Crossing Practice & Timings. Department of Transport and Main Roads report.
Naweed, A. & Larue, G.S. (2015) Somewhere between haste and delay… Investigating the interactions of road users and pedestrians in a dynamic rail level crossing environment. In 19th Triennial Congress of the International Ergonomics Association, 9-14 August 2015, Melbourne, Victoria.
Acknowledgements
The author is grateful to ACRI and the railway partners who funded and helped with the research project, and would also like to gratefully acknowledge the close collaborators of this research: Dr Ross Blackman, A/Prof. James Freeman and Prof. Andry Rakotonirainy.
Dr Grégoire Larue is a research fellow at the Centre for Accident Research and Road Safety – Queensland, Queensland University of Technology (QUT), Australia, and a Research Fellow at the Australasian Centre for Rail Innovation. His contact details are g.larue@qut.edu.au, +61 (7) 3138 4644.