ABSTRACT
Collisions at rail level crossings (RLXs) involving pedestrians represent a significant public safety concern in Australia and internationally. The most recent statistics available show that between 2002 and 2011, 92 pedestrians were struck by trains at RLXs in Australia (Australian Transport Safety Bureau, 2012). In the United Kingdom, over a similar time frame, 72 pedestrians were killed at level crossings (Rail Safety and Standards Board, 2015). The European Railway Agency has reported 373 fatalities associated with collisions at RLXs in Europe in 2012 alone, with approximately 40% of those killed being pedestrians (European Railway Agency, 2014). It has been noted that Australia and the United States have achieved reductions in the numbers of motor vehicle-train collisions, but not in pedestrian-train collisions (e.g., Australian Transport Safety Bureau, 2012; Metaxatos and Sriraj, 2013). To make gains in improving pedestrian safety at RLXs, a new approach is required. Such an approach recognizes that RLXs are complex sociotechnical systems. Taking this perspective, safety at RLXs is the outcome of interactions between social and technical components such as road users, vehicles (road and rail), equipment, and infrastructure. The interactions can be diverse and random, particularly due to the openness
CONTENTS
12.1 Introduction ........................................................................................................................ 135 12.2 The RLX Domain ............................................................................................................... 136
12.2.1 WDA Representation ............................................................................................. 137 12.2.2 Affordances and Constraints at RLXs ................................................................ 138 12.2.3 Design Hypotheses ................................................................................................ 140
12.3 Observations of Pedestrian Behavior .............................................................................. 140 12.3.1 Site Selection ........................................................................................................... 140 12.3.2 Observation Protocol ............................................................................................. 142
12.4 Results ................................................................................................................................. 143 12.4.1 Inter-Rater Reliability ............................................................................................ 143 12.4.2 System State during Observations ...................................................................... 143 12.4.3 Evaluation of Design Hypotheses ....................................................................... 143
12.5 Conclusions ......................................................................................................................... 147 Acknowledgments ...................................................................................................................... 148 References ..................................................................................................................................... 148
of the system with no barriers to system entry in place for many road users including pedestrians and cyclists (i.e., there is no licensing, training, or significant supervision of these users).
