Investigating drivers’ visual search strategies: towards an efficient training interventionTools Konstantopoulos, Panagiotis (2009) Investigating drivers’ visual search strategies: towards an efficient training intervention. PhD thesis, University of Nottingham.
AbstractRoad crashes are the main cause of death of young people in the developed world. The factors that cause traffic crashes are numerous; however, most researchers agree that a lack of driving experience is a major contributing factor. Another reason that has been reported for the increased crashes is that novice drivers have not developed the optimum visual search strategies of their more experienced counterparts. Although several training interventions have tried to improve scanning of novice drivers, they have limited success. The aims of this Thesis are to identify some parameters that influence visual search and to develop an efficient training intervention that will improve drivers’ visual skills. In Experiment 1 an image-based questionnaire was used to assess driving instructors’ and novice drivers’ priority ratings to different areas of the driving scene. Results showed that for both groups the opinions regarding visual field prioritisation were highly consistent when compared to chance. Despite the rating consistencies, group differences were found, across all scenarios with “Rear View Mirrors” being the visual field with the most frequent observed group differences. Certain categories (“Road Ahead” and “Mirrors”) were highly ranked across all scenarios, while other categories were more scenario specific. In Experiments 2 & 3 a novel experimental paradigm was used to investigate the interaction bottom-up and top-down influences upon drivers’ visual attention. Analysis showed that participants’ fixation locations had a stronger relationship with where participants clicked (top down) than with saliency peaks (bottom up). In Experiments 4 & 5 the difference in eye movements between driving instructors and learner drivers was examined during simulated driving. Results showed that driving instructors had an increased sampling rate, shorter processing time and broader scanning of the road than learner drivers. Scenario-specific analysis showed that instructors fixated more than learners on side mirrors while learners showed higher visual allocation to the rear view mirror. It was also found that poor visibility conditions and especially rain decrease the effectiveness of drivers’ visual search. Finally in Experiments 6, 7 & 8 we asked how we can improve learner drivers’ visual skills. Results from Experiments 6 & 7 demonstrated that the ability to distinguish between the eye movements of learner drivers and driving instructors improved as the number of objective differences between the two groups increased across specific scenarios. In Experiment 8 a pilot study showed that a scenario specific training intervention can improve certain aspects of learner drivers’ visual skills. The findings of this Thesis have both theoretical and practical implications regarding drivers’ visual search.
Actions (Archive Staff Only)
|
Tools
Tools
