An investigation into the factors affecting the perception of a train's travelling speed

Abstract

Collisions between cars and trains at railway level crossing junctions continue to occur worldwide, despite efforts to reduce their frequency with educational and practical measures. Many of these collisions occur with car drivers attempting to cross the track in front of an approaching train. The size-speed illusion, first reported by Leibowitz (1985) is a phenomenon whereby larger vehicles appear to move slower than smaller vehicles travelling at the same speed. Clark et al (2013) tested the illusion using laboratory-based experiments, and found that observers routinely underestimated the relative speed of a train, when compared with a smaller vehicle (car). No specific reason for the occurrence of the size-speed illusion has been postulated, but Leibowitz suggested that observer eye movement patterns could be an underlying cause of the illusion. The aim of this thesis was to investigate how observer eye movement patterns influence the size-speed illusion, and consequently the underestimation of a large vehicle’s speed. Experiment 1 tested observers’ perceived judgement of simulated trains and cars approaching in depth in a controlled laboratory setting, with eye movements recorded by an eye tracker. Results confirmed the size-speed illusion and eye movement data showed that patterns of saccades, fixations and smooth pursuit behaviours differed in the case of the longer train, with initial saccades being made to a region further from the front of the train. Experiment 2 and 3 isolated the main types of eye movement patterns that our observers displayed and sought to test whether manipulation of these had an effect on the illusion. Experiment 2 found that manipulating smooth pursuit patterns by placing a target on the front of a long moving object eliminated the illusion. Experiment 3 found that manipulation of fixation and saccadic behaviours with stationary foreground stimuli also reduced the magnitude of the illusion but did not eliminate it entirely. The final experiment trialled three countermeasures designed to replicate the effects of manipulating eye movement patterns as shown in Experiments 2 and 3. Results showed that the intervention based on manipulation of smooth pursuit (alternating flashing lights on the front of a train) was the most effective in reducing the effects of the size-speed illusion. Our results indicate that the use of countermeasures which have the effect of changing eye movement behaviour is most effective in reducing underestimations of a train’s perceived speed, which, hopefully in future, help reduce level crossing collisions.